Publications:

Return to the top of the page

Return to the top of the page

Return to the top of the page

Is there a critical period for novelty exposure-induced enhancement in spatial memory?

Z. Yang, J. Youngblood, A.C. Tang
Abstr. ISDP, 2008

We investigated in the rat how neonatal novelty exposure and juvenile novelty exposure jointly influence cognitive development. During the first 3 weeks of life, one half of each litter was exposed to a non-home cage for 3 min daily while the other half remained in the home cage. During PND 54 to 64 (juvenility), half of the novelty-exposed and half of the home-staying rats were exposed with additional novel experience in a sector open filed. At 5 months of age, offspring spatial memory was evaluated on a moving platform version of the water task, indexed by the amount of time a rat spent in the vicinity of a previous learned escape location 24 hr after the initial training. We found a significant interaction effect between the neonatal and juvenile novelty exposure on the 24 hr spatial memory measure (F(11)=9.948, p=0.009, f=0.951). Follow up tests show that (1) for the rats who did not have juvenile novelty exposure, neonatal novelty exposure enhanced spatial memory in adulthood, which replicates our previous results (t(17)=-2.007, p=0.031, one tailed); (2) for the rats who did not experience neonatal novelty exposure, juvenile novelty exposure also enhanced spatial memory (t(16)=-2.168, p=0.046), which indicates that infancy does not constitute a critical period beyond which novelty exposure would have no effect; (3) for rats with neonatal novelty exposure, Juvenile novelty exposure impaired the spatial memory (t(15)=2.038, p=0.060), which suggests that simple added models are insufficient for explaining the complex interaction between neonatal and juvenile experience.

Two sources of maternal modulations: mother's stress hormone versus mother's behavior

A.C. Tang, Z. Yang, R.D. Romeo, A. Plakio, A. Rosa, A. Gonzales and B.S. McEwen
Abstr. ISDP, 2008

We investigated in the rat (1) whether variations in maternal stress physiology are associated with variations in maternal care behavior and (2) whether the effects of daily 3-min neonatal novelty exposure (NE) are modulated differentially by maternal physiology or maternal behavior. Maternal stress physiology is characterized by basal corticosterone concentration and evoked corticosterone response after 1-min swim stress. Discriminative maternal care behaviors (retrieval order and latency) were characterized separately for the novelty exposed and control rats immediately after the daily NE. Non-discriminative maternal care behaviors (licking and grooming; arched-back nursing) were characterized around the clock during P1-10 (4 times per day throughout the light-dark cycle). Novelty effects on offspring emotionality was assessed in an open field. A mother with a profile of low-basal and high-evoked response tended to show a greater havituation of discriminative care towards the Novel and Home pups (Basal: r=-0.403, p=0.122; Evoked: r=0.599, p=.018) and displayed lower amounts of round-the-clock maternal care (LG-ABN)(Basal: Rs=0.491, p=0.053; evoked: Rs=-0.714, p=.003). A greater novelty effect was observed among offspring whose mothers have a low-basal and high-evoked response profile (basal: Rs=-0.553, p=0.021; Evoked: Rs=0.637, p=0.008). Neither preferential post-NE maternal care (retrieval) nor around the clock measure of LG-ABN showed any correlations with the magnitude of this novelty effect. These results support the conclusion that individual defferences in the amount of maternal care behaviors are unlikely the cause for the reduction of fear by neonatal novelty exposre and instead maternal physiology is a better predictor for early stimulation effect.

Social competitiveness and plasticity of neuroendocrine function in old age: influence of neonatal novelty exposure and maternal care reliability

A.C. Tang, K.G. Akers, Z. Yang, D.P. Delvecchio, B.C. Reeb, R.D. Romeo and B.S. McEwen
Abstr. SFN, 2008

We examined the effects of neonatal novelty exposure and role of maternal care behaviors on late senescent rats' ability to win in social competition and associated changes in the regulation of corticosterone stress response to this social competition. During the first 3 weeks of life, half of each litter received daily 3-min exposures to a novel environment while the other half stayed in the home cage. For the first 10 days, maternal behaviors were observed immediately following neonatal novelty exposure. At 24 months of age, pairs of rats competed against each other for exclusive access to chocolate rewards. We found that novelty-exposed rats won more rewards than home-staying rats (t(10) = 1.82, p < .05), indicating that early experience exerts a life-long effect on this aspect of social dominance. Furthermore, novelty-exposed but not home-staying rats exhibited a habituation of corticosterone release across repeated days of social competition testing (Novel: t(14) = 3.18, p = 0.007; Home: p = 0.411), suggesting that early experience permanently enhances plasticity of the stress response system. Finally, we report a surprising finding that across individual rat families, greater effects of neonatal novelty exposure on stress response plasticity were found among families whose dams provided more reliable, instead of a greater total quantity of, maternal care (r = -0.704, p = 0.016, n = 11). These observations support two conclusions. First, brief exposures to novelty early in life affect: (a) social competitiveness throughout life span, during not only adulthood (Tang et al, 2006, PNAS) but also late senescence; (b) HPA regulation, specifically increasing plasticity of corticosterone stress response as a function of experience. Second, the amount of enhancement in HPA plasticity differs between different rat families, with the greatest enhancement found among litters that received more reliable maternal post-stimulation maternal care. The total amount of post-stimulation maternal care did not appear to mediate this within-litter neonatal novelty exposure effect.

When more stress hormone is good: Reduced interference in a Stroop task is associated with higher salivary cortisol

M.T. Sutherland, Z. Yang and A.C. Tang
Abstr. SFN, 2008

Stress is known to modulate cognitive performance, and differences in stress hormone regulation may account for variability observed during cognitive task performance. Using data from a Stroop experiment designed for another purpose, we investigated whether an individual's ability to resist interference from task-irrelevant, yet salient information is in part influenced by the state of the stress response system and whether anterior cingulate cortex (ACC) activity, known to be sensitive to such interference, is similarly influenced. Participants were asked to respond with a button-press to indicate the color of color-word stimuli. In this task, when color and word meaning are congruent (C: "RED" printed in red color), interference from word meaning is not present. When incongruent (I: "GREEN" printed in red), the subject must overcome the interference from word meaning ("GREEN") and respond to the color (red). To measure individual differences in the ability to overcome interference, a difference score between I and C trial reaction times (D-RT, interference effect) was calculated. Individual differences in autonomic activity were indexed by a single salivary cortisol measure obtained at the end of the task. To assess differential neural activity in the absence and presence of interference, first, high-density EEG data was collected and processed with a blind source separation algorithm (second-order blind identification: SOBI) to extract frontal midline activity thought to originate from ACC. Then, the amplitude of a response-locked negative deflection in the ERP peaking ~50ms after the response, the correct-related negativity (CRN), was quantified. Finally, a difference score between I and C trial CRN amplitude (D-CRN) was computed. We found that participants with higher cortisol were better able to minimize interference from task-irrelevant information as indicated by a negative correlation (r [19] = -0.701, p = 0.001) between the salivary cortisol measure (cubic root transformed) and D-RT score (log transformed). Furthermore, participants with higher cortisol also showed increased CRN sensitivity to the presence of interfering information as indicated by a marginally significant correlation (r [19] = -0.416, p = 0.068) between cortisol and D-CRN score (cubic root transformed). These results suggest that individual differences in the self-regulation of cortisol are related to ones ability to overcome interference from salient, yet task-irrelevant information. To the extent that the CRN measure assessed in the present task tracks ACC functioning, these results suggest a relationship between ACC responsivity to interference and cortisol.

Beyond maternal care: modulation of early stimulation effects by maternal physiology

Z. Yang, R.D. Romeo, B.S. McEwen and A.C. Tang
Abstr. SFN, 2008

Across species, work of Rutter, Bowlby, Harlow, and Meaney repeatedly confirm a common wisdom that maternal care behaviors are of paramount importance to the offspring's development. Here in the context of an early stimulation experiment, we show that maternal physiology are strongly associated with maternal care behavior, thus providing a possible cause influence over offspring psychological development. Using the neonatal novelty exposure, we exposed half of each litter of rats daily to a relatively novel non-home cage (Novel group) for 3 min and the other half remained in the home cage (Home group) during postnatal days 1-21. Dams showed preferential care towards Home pups with the Home pups being the dams' first retrieval choice more often than Novel pups (t(17)=2.426, p=0.027) and being retrieved sooner than Novel pups (t(15)=-1.818, p=0.045, one-tailed). This maternal discrimination in her distribution of care was related to her stress response profile. The mothers with lower CORT and higher evoked CORT showed a decrease in her preferential care toward the Home pups across the daily novelty exposure (Basal: r=-.403, p=.122; Evoked: r=.599, p=.018), suggesting that discriminative maternal care behaviors are in part controlled by her stress response system. Influence of maternal physiology on her care behaviors also applied to the round-clock non-discriminative maternal care measures. The higher the maternal basal CORT and lower the evoked CORT, the more the dams engaged in arched back nursing (Basal: Rs=.499, p=.049; Evoked Rs=-.690, p=.004). Most importantly, despite a lack of preferential maternal care, Novel rats were less fearful than the Home rats when being placed in a novel environment during juvenility [F(1,18)=6.421, p=.021, f=.597]. The magnitude of this novelty-exposure across different rat families was modulated by maternal physiology, with the maternal basal CORT negatively (Rs=-.553, p=.021) and evoked CORT positively (Rs=.637, p=.008) predicting novelty effect. These findings indicate that individual differences in maternal care are associated with individual differences in maternal stress response profile. Thus offspring developmental differences previously thought to be caused by a greater maternal care may be caused instead by differential maternal stress physiology. Furthermore, the mothers stress response profile can set the magnitude and direction of other environmental influences, thus jointly shaping offspring development. We thank B. Reeb and D. Delvecchio for assisting with novelty exposure and A. Plakio, A.R. Silva-Banuelos, and A. Gonzales with coding and data entry.

Predicting adult stress response by neonatal events

Z. Yang, R.D. Romeo, A.M. Korzekwa, B.C. Reeb, B.S. McEwen and A.C. Tang
Abstr. Brain Research, 2008

Using the rat as a model system, we investigated how adult physiological stress response to physically painful and psychologically stressful events is jointly modulated by temporally remote early life events and temporally immediate events. During postnatal days 1-21, half of each litter of rats were exposed to a relatively novel non-home cage (Novel) for 3 min/day and the other half remained in the home cage (Home). This neonatal treatment has been shown to induce long-lasting changes in the corticosterone stress response and enhancement in social and cognitive functions. At P24, recovery from behavioural inhibition in a novel environment was measured by a disinhibition score, defined as an increase in activity between two 20s exposure to an open field. At 14 months of age, immediately before repeated blood sampling (3, 15, 30, 60, and 90 min) via tail nicking, half of the Novel and Home rats were primed by a brief exposure to a novel environment (2 min in open-ended plastic bottle) while the other half waited in the home cage. The repeated blood sampling constitutes a sequence of physically painful and psychologically stressful events. We found: (1) Open-field disinhbition at 24 days of age predicted average CORT at 14 months of age (Rs=.366, p=0.047, n=30); (2) average CORT was jointly determined by the temporally recent and remote novelty experience [ F(1,26)=5.01, p=0.034]. Specifically, the 2 min brief novelty exposure induced a potentiation in CORT output among the Home rats(Home-NR: 352?18; Home-R: 410?21) and an attenuation among the Novel rats (Novel-NR: 410?21, Novel-R: 383?24). Together these results suggest that history of stress experience can transcend time to create differential physiological response to the same physical events via complex interactions.

Who performs better under interference? Prediction from stress response profile

Z. Yang, M.T. Sutherland and A.C. Tang
Abstr. Brain Research, 2008

Individuals differ in their ability to focus on the task while overcoming the influence of task irrelevant yet salient sensory information. Here we test the hypothesis that individual differences in stress response profile predict individual differences in ability to overcome interference. One critical aspect of this profile is the responsiveness of the HPA axis to a salient stressor. We measured HPA responsiveness on three testing days separated by one week interval using the Triers Social Stress Task (TSST), which has been shown to reliably evoke an increase in saliva cortisol. Individual differences in interference resistance was measured using the Stroop interference effect, defined as the reaction time differences between responding to a congruent stimulus (e.g. word RED printed in color red) and an incongruent stimulus (e.g. word RED printed in color green). As the subject was instructed to respond according the printed color instead of word meaning, the meaning of the word interferences with task performance. Subjects performed the color Stroop and the TSST tasks on different days of laboratory visits and saliva samples were collected both immediately before and after the TSST task for measuring cortisol. HPA responsiveness was measured three times by a TSST-related increase in saliva cortisol. Preliminary findings show that the greater the evoked cortisol response to TSST, the less one's performance in the Stroop task suffers from the interference effect (TSST1: Rs=-0.771, p = 0.036, n=6; TSST2 Rs=-0.600, p=0.104, n=6; TSST3: Rs=-0.800, p=0.052, n=5). These results suggest that even in a task context that lacks apparent stressors, cognitive performance may nevertheless be regulated by one's stress response characteristics.

When more stress hormone is good: reduced interference in a stroop task associated with higher salivary cortisol

M.T. Sutherland, Z. Yang, and A.C. Tang
Abstr. Brain Research, 2008

Differences in stress hormone levels may account for variability in task performance. Using data from a Stroop experiment, we investigated whether an individual's ability to resist interference from task-irrelevant, yet salient information is in part influenced by the state of the stress response system and whether anterior cingulate cortex (ACC) activity, known to be sensitive to interference, is similarly influenced. Participants were asked to make button-presses to indicate the color of color-word stimuli. In this task, when color and word meaning are congruent (C: "RED" printed in red color), interference is not present. When incongruent (I: "GREEN" printed in red), interference from word meaning ("GREEN") must be overcome and a response to color (red) given. To measure individual differences in the ability to overcome interference, a difference between I and C trial reaction times (D-RT, interference effect) was calculated. To assess neural activity in the absence/presence of interference, the amplitude of a response-locked ERP deflection peaking ~50ms after the response (correct-related negativity, CRN), was quantified. A difference between I and C trial CRN amplitude (D-CRN) was computed. Individual differences in stress reactivity were indexed by post-task salivary cortisol. Participants with higher cortisol concentration were better able to minimize interference from irrelevant information as indicated by a negative correlation (r[19]= -0.701, p=0.001) between cortisol (cubic root transformed) and D-RT (log transformed). Participants with higher cortisol concentration also showed increased CRN sensitivity to the presence of interfering information as indicated by a marginally significant correlation (r[19]= -0.416, p=0.068) between cortisol and D-CRN (cubic root transformed). These results suggest that individual differences in cortisol are related to one's ability to overcome interference from salient, yet irrelevant information. To the extent that the CRN measure tracks ACC functioning, these results suggest a possible relationship between ACC responsivity and cortisol.

Enhancement of social competitiveness among senescent rats by neonatal experience of novelty

D.P. Delvecchio, Z. Yang, K.G. Akers, and A.C. Tang
Abstr. APS, 2008.

Abstract: We show that in the rat, neonatal novelty exposure results in a permanent enhancement of social competitiveness. Aged rats (24 months of age) were trained to compete for chocolate rewards against another rat. Rats with early experience of novelty won more rewards than their siblings who stayed at home.

Supporting Summary: Neonatal stimulation affects cognitive function throughout life span (Meaney et al, Science 1988; Tang, Learn & Mem 2000). Recently this effect has been extended to the domain of social function during adulthood (Tang et al, J. Neurosci. 2003; PNAS 2006). Here we investigate whether the induced modifications of social function can persist into senescence. Specifically, we examined the effect of neonatal novelty exposure on social competitiveness using a chocolate competition task.

During the first 3 weeks of life half of the pups were exposed to a non-home cage for 3min per day (Novel), while the other half of their siblings remained in the home cage (Home). Both groups were handled by the experimenter equally and separated from their dams for less than 15 min. At 24 months of age, rats were trained to obtain chocolate individually for 5 consecutive days to reach asymptotic performance. Pairs of Novel and Home rats then competed to obtain chocolate for two consecutive days. Pairs were formed pseudo randomly with the following constraints. Within pairs, first the rats must come from the same litter to enable within-litter between-sibling analysis and second the final training performance and body weight between the Novel and Home rats must not differ significantly to insure that learning performance and weight differences were not confounding factors.

The number of winning trials was used as index for social competitiveness. Winning patterns across the two days of competition differed between the Novel and Home rats (Novelty by Day interaction: F(1,10) = 8.92, p = 0.014). On Day 1, when the testing situation was surprising due to the unexpected presence of a competitor, Novel rats won significantly more rewards than Home rats (t(10) = 1.82, p = 0.0495). This competitive advantage was unlikely caused by a difference in their speed of reaching the rewards or by a difference in motivation, as latencies on the testing days did not differ between groups (Novel: 5.95 +/- 3.12 s; Home: 6.60 +/- 3.46 s; p=0.611). On Day 2, when the testing situation was no longer surprising, Novel and Home rats did not differ (p = 0.336).

These results replicated our previous finding of enhanced social competitiveness during adulthood (Tang et al PNAS 2006) and extended the finding into senescence. The fact that this enhancement is only expressed when the competitive situation is novel suggests that one origin of individual differences in social competitiveness may be the individual's response to novelty. Indeed, aged Novel rats from the same cohort showed less behavioral inhibition towards a novel odor than the aged Home rats (Yang et al 2008, APS abs). These results demonstrate that minutes of daily exposures to a non-home environment early in life can have a permanent effect on adult social competitiveness that persists into senescence. See companion abstract for a permanent reduction behavioral inhibition towards novelty by Yang et al. APS, 2008.

Early experience permanently reduces behavioral inhibition towards novelty

Z. Yang, A. Plakio and A.C. Tang
Abstr. APS, 2008.

Abstract: We show that in the rat, neonatal novelty exposure results in a permanent reduction in behavioral inhibition to novelty. Aged rats (26-months of age) were exposed to a novel odor. Rats with early experience of novelty approached the novel odor more frequently and sooner than rats who stayed at home.

Supporting Summary: Using the rat as a model system, we (Tang et al., PNAS 2006) have found that seemingly mild neonatal stimulation can result in enhanced spatial and social memory during adulthood. Others (Meaney et al Science 1988) have found that similar early life experience can result in attenuated cognitive aging among aged rats. Here we address the question of whether brief exposures to novelty early in life can have persistent impact on response to novelty among aged rats.

In the present study, we followed 39 pups from birth to 26 months of age. During the first 3 weeks of life (P1-21), half of the male pups were exposed to a non-home cage for 3min per day (Novel, N=19) and the other half remained in the home cage (Home, N=20). Novel and Home rats were both separated from their dams briefly for less than 15 min and the contact with experimenters were matched during this novelty exposure procedure. At 26 months of age, their responses to a novel odor were measured during a 5-min session in a testing cage containing a 7.5 by 9 by 5.5 inch cardboard box with an odor bottle inside.

Based on previous findings from open field tests and habituation experiments, we expected that Novel rats were more explorative and less fearful in the face of novelty and habituated more quickly to novel stimulus. Three behavior measures were analyzed here: (1) frequency of approaching the novel odor box defined as number of times that the rat opened the door of the novel odor box or enter the novel odor box; (2) the latency to approach the box containing novel odor; (3) total time spent in the box containing the novel odor. We found that (1) Novel rats engaged in novel odor approach behaviors more often than Home rats (t(35)=2.655, p=.006); (2) The latency for Novel rats to approach the Novel odor is marginally shorter than Home rats (t(35)=1.661, p=.053); and (3) Novel rats spent less time in the novel odor box than Home rats (t(35)=1.828, p=.039).

The increased frequency and reduced latency in approaching the novel odor among the novelty exposed rats are similar to the increased frequency and reduced latency towards a novel object observed among monkeys who were exposed to a non-home cage repeated when they were young (Parker et al, J. Psyneuen., 2007). Taken together, these results support the notion that early exposure to novelty can permanently reduce behavioral inhibition towards novelty throughout an individual's life span. See companion abstract for a permanent enhancement of social competitiveness among the Novel rats by Delvecchio et al. APS, 2008

Between-sibling and between-family difference in body weight: the role of neonatal novelty exposure and maternal stress response

Z. Yang, B.C. Reeb and A.C. Tang
Abstr. ISDP, 2007.

In a longitudinal study, we investigate how mother rats' individual differences in stress response account for between-family (litter) and between-sibling differences in her offspring' body weights. To create between sibling differences, we used a split-litter design in which half of each litter was exposed to a non-home cage for 3 min per day and the other half of their siblings remained in the home cage during the first 3 weeks of life. Analysis of covariance revealed that (1) concerning between-sibling differences, at both 3 weeks and 13 months of age, although novelty exposure did not show consistent effect on body weights across all litters, it showed significant interaction with maternal stress response----the faster the mother was able to mount a corticosterone response to a one-minute swim challenge, the greater the novelty-induced between-sibling difference in body weight; in contrast, at the onset of adulthood (100 days of age) and senescence (19 months of age), both of times of major developmental change, this maternal modulation of novelty effect was absent; (2) concerning between-family differences, the correlation between maternal stress response and the average litter weight was not significant at any of the four points during development. These results indicated that neonatal novelty exposure induces between-sibling differences in body weight depending on the mother rat's stress response profile but only at some points during development and that there is no evidence that maternal individual differences in stress response profile can account for between-family differences in offspring body weight.

Dissociating the effects of neonatal novelty exposure and maternal stress response on adult offspring spatial memory

A.C. Tang, Z. Yang, A.M. Korzekwa, B.C. Reeb, R.D. Romeo and B.S. McEwen
Abstr. ISDP, 2007.

In a longitudinal study of 104 rats from 22 litters, we investigate how neonatal novelty exposure and the mother rats' individual differences in stress response profile jointly determine between-sibling differences in cognitive development. One half of each litter was exposed to a non-home cage for 3 min daily during the first 3 weeks of life (Novel) while the other half remained in the home cage (Home). Offspring spatial reference memory was indexed by the amount of time a rat spent in the vicinity of a previous escape location 24 hr after the initial escape training. Reversal learning was indexed by how quickly the rat reversed to a new escape location after a single experience of the new location. We found: (1) at both 4 weeks and 9 months of age, Novel rats showed both a greater reference memory and greater reversal learning than the Home rats; (2) at 4 weeks of age, these novelty effects were modulated by maternal stress response profile---the faster the mother was able to mount a corticosterone response to a one-minute swim challenge, the greater novelty-induced enhancement was observed in her offspring; (3) long after the rats entered adulthood at 9 months of age, novelty effects persisted while maternal modulation of these effects ceased to exist. This temporal dissociation indicates that neonatal novelty exposure, although brief in comparison to the mother rats' continuous presence, outlasts maternal influence and maintains its positive impact long into adulthood.

Testing activation-timing dependent plasticity in human cortical networks

M. Nakazawa, Z. Cui, N. Furl, R.J. Dolan and A.C. Tang
Abstr. SFN, 2007.

At the microscopic level, spike-timing is critical for the induction and direction of the plasticity. This lower level plasticity gives rise to and underlies changes at the macroscopic level. Here using magnetoencephalography (MEG), we test an analogous hypothesis at the level of cortical networks that activation-timing of cortical areas can be critical for the induction and direction of cortical plasticity. We used median nerve stimulation of the left and right wrists to activate two feed-forward pathways that converge onto the left and right primary somatosensory cortices (L-SI, R-SI) and subsequent high level processing areas. Several patterns of stimulation were used to investigate the effect of activation timing: (1) simultaneous R and L stimulation; (2) L precedes R (LR); (3) R precedes L; (4) random sequence of single L and R stimulation. We combined a blind source separation algorithm, second-order blind identification (SOBI), with BESA to extract continuous time course of activation of specific cortical areas from MEG data (e.g. L-SI, R-SI). From these time courses, we constructed somatosensory evoked fields (SEF) for single L and R stimulation and we then compared these SEFs obtained before and after the 3 paired stimulation blocks and the control block. We will report differential changes in the SEFs resulting from different activation-timing conditions. We will also compute changes in effective connectivity among different cortical regions using Granger causality.

Limited modulation of Stroop interference effects by up to 3 preceding trials

M.T. Sutherland and A.C. Tang
Abstr. SFN, 2007.

In a color-naming Stroop task, the two dimensions of a stimulus can be either compatible (C) or incompatible (I) (C: the word "RED" in red hue; I: the word "RED" in blue hue). The reaction time (RT) difference between I and C trials (the interference effect) has been shown to depend upon the compatibility of the preceding trials (e.g., Kerns et al., Science, 2004). Evidence for the influence of preceding trials beyond the immediately preceding trial comes from a study employing a continuous pseudo-random sequence of I and C trials in a flanker task (Durston et al., NeuroImage, 2003). In that study, only trials with the form cn-I and in-I (lower case letters represent preceding trials, upper case letters: current trials, and n: the number of preceding trials) were assessed. Increasing numbers of c or i trials before an I trial differentially affected both RT and activation in prefrontal brain regions. Here, we designed a blocked experiment to more systematically examine whether RT on an end-trial was affected by the number of preceding-trials and if it is, how many preceding trials have an influence. Twenty subjects performed a Stroop color identification task requiring button press responses. Each trial block consisted of two parts, the priming-trials (pn: p = [c, i], n = number of trials) and the end-trial (e = [I, C]). Within each of the four block types, cn-I, in-I, cn-C, and in-C, n was varied from 1, 3, 5, to 7. To better isolate the effects of priming sequence length, we used an inter-block interval (3000ms) which was noticeably longer than the inter-trial interval with in a block (1500ms) to "reset" the neural system to a relatively similar initial state. End-trial RT was not only jointly influenced by the compatibility of the end-trial and priming-trials, but by priming sequence length as well (3 way interaction: F[3, 57] = 17.04, p < 0.001, ? = 0.47). Furthermore, while RT on end-trials did not differ among sequence lengths of 3, 5, and 7, RT from the priming sequence length of 1 differed significantly from those with lengths of 3, 5 and 7 (F(1, 57) = 45.08, p < 0.001). These results suggest that, (1) the length of the priming sequence influences end-trial RT, and (2) a limited window size of 3 priming-trials affects end-trial performance. To the extent that the end-trial RT reflects activity in brain regions that serve a conflict monitoring function, these results lead to a working hypothesis that dorsal anterior cingulate activity would show similar joint modulation by the compatibility of the priming- and end-trials and the sequence length of priming-trials. We will test this hypothesis by simultaneously collecting high density EEG and behavioral data during the above described task.

Measuring Feed-forward and Feed-back Influence in the Brains of 2-Day Old Human Infants by Integrating SOBI and Granger Causality

A. Korzekwa, P. Sun, P.G. Grieve, J.R. Isler, M.M. Myers, W.P. Fifer and A.C. Tang
Abstr. SFN, 2007.

By combining a blind source separation algorithm called Second-order blind identification (SOBI) with Granger causality analysis, feed-forward and feed-back influences between the frontal and visual cortices can be measured in adult humans non-invasively from high density EEG (128 channels) data collected while the subjects are at rest or performing free viewing (SFN 2007 Abstract). Here we applied the same techniques to the analysis of high density EEG data (128 channels) collected from 2-day old human infants during sleep. First, the raw continuous EEG data were screened for large artifacts that may be associated with electrode movement due to changing sleep positions. Second, SOBI was applied to segments of continuous EEG while respecting the boundaries between clean and contaminated data segments (i.e. disallowing computation of cross-correlation across the boundaries). Third, the sensor space projections of all SOBI components were generated and neuronal components were identified based on the topography of the 2D projection maps. Fourth, from this set of neuronal components, two components of interest were identified: 1) a component capturing focal frontal midline activity, and 2) a component with focal posterior activation symmetric to the midline. Finally, Granger causality, a quantity that measures directional influence between the two brain locations is computed using the time courses from the frontal and posterior SOBI components. Preliminary analysis shows that SOBI can extract multiple neuronal components with distinct topographies from continuous EEG data collected from sleeping infants at 2-days of age despite the presence of large artifacts. This result indicates that SOBI can recover the continuous time course of activity from distinct brain regions using on-going EEG (i.e. without event-related potentials) and without requiring the subjects to perform any tasks. When combined with Granger causality analysis, this novel technique may enable the tracking of directional influences between any two brain regions throughout development. The task-independence of this approach also enables the monitoring of individuals who are unable to cooperate due to age, state, or a disease condition.

Determining Gender from Local Network Synchronization in the Frontal Cortex

P. Sun, Z. Yang, A.M. Korzekwa, M.T. Sutherland, M. Nakazawa and A.C. Tang
Abstr. SFN, 2007.

We investigated the possibility of telling male and female brains apart by examining the dynamics of brain activity in the frontal cortex within a small time window. Continuous high density EEG data were collected while the subjects (8 males and 8 females) experienced the following conditions: a. eyes-closed; b. eyes-open; c. eyes-open viewing a nature video; d: imaging visually the scene in c; e: listening to the audio part of the video. First, we used a blind source separation algorithm called Second-order blind identification (SOBI) to extract a neuronal component that corresponds to activity from a local region within the frontal cortex. Second, we used a support vector machine (LS-SVM) to classify males and females using the power in the theta, alpha, beta, and gamma bands. The time series of each subject's frontal component was truncated into short overlapping windows (10 sec window with 8 sec overlap), and for each window, the power level for each of the frequency bands was calculated. These four quantities set up the feature space for the classification problem; each window produces a 4-dimensional vector to feed into the SVM. We adopted the Radial Basis Function (RBF) as the kernel of the SVM, which is nonlinear and maps the original 4-dimensional feature space to infinite dimensional space. During the training phase of the SVM, 10% of the data from each subject was used to make up the training set and the remaining 90% as the testing data set. Cross-validation within the training set was used to search for the optimal regularization and bandwidth parameters. Three experiments were carried out: (1) training and testing using data from all conditions and using all 4 features; (2) training and testing using data from only the eyes-closed condition and using all 4 features; (3) training and testing using all data and using all subsets of the 4 features. We found: (1) with the entire time series and all 4 band power indexes, an accuracy of 80% was achieved, (2) with the time series from a single condition and all 4 band power indices, an accuracy of 91% was achieved (this improvement in accuracy benefits from the reduced variation of the frequency patterns across experimental conditions); (3) the highest classification accuracy was always achieved with all four band of indices. A classification accuracy of over 90% is remarkable, particularly given that only 10% of 5 min data was used for training the classifier, only 4 features were used to represent the data, only 10 sec of data were used for each sample, and only activity from one location within the brain was used.

Feed-forward versus Feedback Processing in the Human Brain: Distinct Directional Influences Revealed by Integrating SOBI and Granger Causality

A.C. Tang, M.T. Sutherland, P. Sun, Y. Zhang, M. Nakazawa, A.M. Korzekwa, Z. Yang and M. Ding
Abstr. SFN, 2007.

Feed-forward and feedback processing are two distinct yet highly interactive modes of neural activity underlying normal and abnormal human cognition. Here we characterize the dynamic processes that contribute to these two modes of cognitive operation. We used a blind source separation algorithm called Second-order blind identification (SOBI) to extract from high-density scalp EEG (128 channels) two components that index neural activity in two distinct local networks: one in the frontal and the other in the visual cortex. We then applied Granger causality analysis to the SOBI-recovered neuronal signals from these two local networks to characterize feed-forward and feedback influences between them. Continuous EEG data were collected while the subjects experienced the following conditions: a. eyes-closed; b. eyes-open; c. eyes-open viewing a nature video. Subjects were allowed to freely move their eyes during viewing and to blink whenever needed. No behavioral responses were required. To characterize stability as well as variations across repeated measures within a given individual, each subject was tested at Week 0, Week 1 and Week 4+. For all subjects, we were able to repeatedly identify a frontal and posterior component across all 3 repeated observations. There was no significant difference in the precise locations of these components from one observation to another. This result shows that without using any specific event markers, as in ERP analysis, SOBI can reliably recover time course of activation from the same regions of the frontal and visual cortices with inter-observation intervals of weeks. Analysis of local network synchrony showed that the frontal and visual cortex activations are characterized by a prominent theta and alpha peak, respectively, and theta power increases relative to the alpha power as the amount of visual processing increases across conditions a, b, and c. Analysis of the long-range network synchrony between the frontal and visual cortices showed that feed-forward influence is dominated by alpha while feedback influence is dominated by theta band activity. This direction-selective dominance pattern is jointly modulated by situational familiarity and demand for visual processing, with situational familiarity leading to a decrease in feed-forward theta dominance and the visual processing demands leading to an increase in feed-forward theta dominance. These findings demonstrate that combining SOBI and Granger causality enables a non-invasive and non-intrusive measurement of directional influence between two functionally distinct brain regions under free viewing conditions. (Supported by a grant to ACT from the Sandia National Laboratory)

Priming Adult Stress Response by Neonatal Novelty Exposure

Z. Yang, R.D. Romeo, M. Nakazawa, A.M. Korzekwa, B.C. Reeb, B.S. McEwen and A.C. Tang
Abstr. SFN, 2007.

Using the rat as a model system, we investigated how adult physiological stress response to physically painful and psychologically stressful events is jointly modulated by temporally remote early life events and temporally immediate events. During postnatal days 1-21, half of each litter of rats were exposed to a relatively novel non-home cage (Novel group) for 3 min/day and the other half remained in the home cage (Home group). During this daily novelty exposure, the mother rats were separated from both the Novel and Home pups and both groups were touched equally by the experimenter. This neonatal treatment has been shown to induce long-lasting changes in the corticosterone stress response and enhancement in social and cognitive functions. At 14 months of age, half of the Novel and Home rats were exposed to a brief (2 min) restraint while the other half waited in the home cage (a 2 x 2 design). At 3, 15, 30, 60, and 90 min after the onset of the restraint, blood samples were collected from each rat by tail nicking (note that each blood drawing entailed ~ 4 min of restraint). This repeated blood sampling constitutes a sequence of physically painful and psychologically stressful events. We found that the average CORT levels across all sampling points is jointly modulated by the temporally recent (2 min restraint immediately before) and temporally remote (novelty exposure 14 months before) events [Novelty by Restraint interaction, F(1,26)=5.01, p=0.034, effect size = 0.45]. Specifically, the 2 min brief restraining prior to blood collection had opposite effects for the Novel and Home rats with Home rats showing a restraint-induced potentiation in CORT output (ng/ml) (Home-NR: 352 18; Home-R: 410 21) and Novel rats showing a restraint-induced attenuation (Novel-NR: 410 21, Novel-R: 383 24). This pattern of interaction indicates that very brief neonatal exposures to novelty can result in a reversal of the priming effect by a preceding stressor on the responses of the HPA axis to subsequent stressful events. We suggest that it is via a sequence of such higher-order effects, differences in early life stress experience transcend time to bifurcate the physiological meaning of an adult stressful event for different individuals. (We thank M. Sutherland, J. Rice, and A. Dumars for their help with blood collection).

Explaining Early Stimulation Effect: Maternal Modulation as an Alternative to Maternal Mediation

B.C. Reeb, R.D. Romeo, B.S. McEwen, and A.C. Tang
Abstr. SRCD, 2007.

The maternal mediation hypothesis has been used to account for the effects of neonatal stimulation on offspring physiology and behavior (Liu et al., 1997). Recent primate studies support an alternative hypothesis of stress inoculation (Parker et al., 2006), challenging the maternal mediation hypothesis. Here we tested a maternal modulation hypothesis which states that early stimulation has a direct effect on the pups' physiology and behavior and the magnitude as well as direction of this stimulation effect is modulated by the dams' post-stimulation behavioral patterns (Jones & Tang, 1999). Specifically, early stimulation-induced improvements in cognition and emotional reactivity should be more likely observed when the mother behaves as a stress attenuator and, conversely, functional impairments may result when the mother behaves as a stress potentiator. One suggested potentiator is a high variability or low reliability in post-stimulation maternal care (Jones & Tang, 1999).

To create a source of neonatal environmental difference that is orthogonal to maternal individual differences, we used the neonatal novelty exposure procedure (Tang, 2001) in which half of each litter was subject to early stimulation while the other half remained in the home cage from postnatal days (PND) 1-21. Specifically, from a total of 22 litters, 53 male rat pups were exposed to a novel cage for 3 min daily (Novel group) while the 53 littermate controls remained in the home cage (Home group). Early stimulation effects were assessed in an open field task on PND 24 and a working memory version of the Morris water task on PND 32-37. From PND 1-10, variability of maternal licking and grooming during the first 5 min immediately following reunion with the pups was measured as an index of reliability in maternal care. In addition, the dams' circulating basal and evoked corticosterone responses measured 5-6 days after weaning was used as an index of individual differences in the maternal trait stress response. Both maternal care variability and stress response were used as predictors for early stimulation effects on emotional and cognitive development.

Within-litter comparisons revealed that, independent from maternal individual differences, Novel rats showed greater disinhibition in the open field [F(1,20)=4.124, p=0.028] and better spatial working memory than Home rats [F(1,21)=5.254, p=0.032]. The magnitude and direction of the novelty effect varied from litter to litter, with the majority of Novel rats within the litters showing increased disinhibition and enhanced memory and a few showing decreased disinhibition and impaired memory when compared to Home rats. Both disinhibition and spatial working memory were significantly correlated with maternal baseline CORT (disinhibition: rs=-0.457, p=0.047; spatial memory: rs=-0.530, p=0.013) and evoked CORT responses (p's<0.05). Variability in post-stimulation maternal licking and grooming was negatively correlated with the within-litter novelty effects (spatial memory: rs=-0.435, p=0.043; disinhibition: rs=-0.405, p=0.086). The current data suggest that neonatal novelty exposure induces a change in the offspring's emotional and cognitive development and that maternal individual differences serve to modulate the direction and magnitude of this novelty-exposure induced effect, thereby offering support for a modulating instead of a mediating role for maternal influence.

Creation of between-sibling differences: an induction by 60-minute neonatal non-shared environment

A.M. Korzekwa, Z. Yang, B.C. Reeb and A.C. Tang
Abstr. APS, 2007.

In rats, between-sibling cognitive differences were observed at 9-months of age, long after exposing half of the rat pups within a litter to 3-minutes daily neonatal novelty exposures during the first 3-weeks of life. This finding may offer insight regarding why children from the same family can be so different.

Impact of Mother Rats' Physiological Stress Response on Adult Offspring's Body Weight

Z. Yang, B.C. Reeb and A.C. Tang
Abstr. APS, 2007.

Offspring of rat mothers who show a rapid stress response tend to weigh more than those whose mothers have a sluggish stress response. This maternal modulation of physical development was only observed among siblings that experienced 3-min daily exposures to novelty during the first 3 weeks of life.

Early Life Modulators of Adult Social Engagement during Dyadic Interaction

A.C. Tang, K.G. Akers, A.M. Korzekwa and B.C. Reeb
Abstr. APS, 2007.

Using a novel framework for decomposing social interaction, we show that adult ability to engage a conspecific in social interaction is modulated by early life events, with the repeated neonatal novelty exposure leading to an increase and a single hypoxia event leading to a decrease in social reciprocation received.

Creating Temperamental Differences between Sibling Rats via 60-Minutes of Non-Shared Environment

M. Nakazawa, B.C. Reeb, N.M. Vandiver and A.C. Tang
Abstr. APS, 2007.

Exposing a subset of rat pups from a litter to a non-home cage for 3-min daily during the first 3-weeks life created between-sibling differences in novelty response during adulthood. This temperamental difference can be revealed, rather surprisingly, by the first few seconds of exposure to a novel environment.

What makes an adaptive rat - "better" mothers or early exposure to novelty?

A.C. Tang, K.G. Akers, B.C. Reeb, A.M. Korzekwa and L.N. Rogers
Abstr. ISDP, 2006.

Mild neonatal stimulation via brief exposures to a relatively novel non-home environment during infancy is known to induce functional enhancement throughout the lifespan. Here, we critically test the hypothesis that preferential maternal care is the mediator of such enhancement. In Experiment 1, we exposed half of each litter to a novel cage for 3 min daily while the other half remained in the home cage during PND1-21. Preferential maternal care was indexed by retrieval latency measured immediately after novelty exposure. Home-staying pups were retrieved sooner than novelty-exposed pups (p=0.03), suggesting that dams gave lower priority of maternal care to novelty-exposed pups. In Experiment 2, we determined the relationship between retrieval priority and level of maternal care behavior (licking-grooming and arched-back nursing) during PND1-10. A high retrieval priority (shorter retrieval latency) was associated with a high level of maternal care (p=0.05), supporting the use of retrieval priority as an indicator for maternal care. In Experiment 3, we investigated the consequence of differential maternal care behavior in cognitive and social functions when rats in Experiment 1 became adults. Despite the unfavorable maternal care received during infancy, novelty-exposed rats displayed enhanced spatial working memory (p=0.04) and increased success in social competition for resources (p=0.05) relative to controls. Together, these findings lead to a surprising conclusion that preferential maternal care is neither necessary nor sufficient for early stimulation-induced functional enhancement.

Explaining early stimulation effect: maternal mediation or maternal modulation?

B.C. Reeb, R.D. Romeo, B.S. McEwen, and A.C. Tang
Abstr. ISDP, 2006.

The maternal mediation hypothesis has been used to account for the effects of neonatal stimulation on offspring physiology and behavior (Liu et al., 1997). Recent primate studies support an alternative hypothesis of stress inoculation (Parker et al., 2006), challenging the maternal mediation hypothesis. Here we test a maternal modulation hypothesis that aims to integrate the role of direct stimulation effect and maternal behavior. This hypothesis states that early stimulation has a direct effect on pups' physiology and behavior and the magnitude as well as direction of this stimulation effect can be modulated by maternal individual differences in stress response. To create a source of neonatal environmental difference that is orthogonal to maternal individual differences, we used a split-litter design in which half of each litter (Novel) was subject to early stimulation (3 min exposure to a novel cage) while the other half (Home) remained in the home cage from postnatal days 1-21. Within-litter comparisons revealed that, independent from maternal individual differences, Novel rats had better spatial memory than Home rats (p=0.032). The magnitude and direction of the novelty effect varied from litter to litter, with the majority of litters showing enhanced and a few showing impaired memory. This variation was accounted for by both maternal baseline CORT and evoked CORT responses (significant correlations, p's < 0.05). These results suggest that neonatal novelty exposure induces an enhancement in the offspring's cognitive development and that maternal individual differences serve to modulate the direction and magnitude of this novelty-exposure induced effect.

"Social control" between two strangers - impact of neonatal novelty exposure on adult social interaction

K.G. Akers, B.C. Reeb, R.D. Romeo, B.S. McEwen, J. Kagan, R.J. Dolan, and A.C. Tang
Abstr. ISDP, 2006.

Infant pups that spent 3 minutes a day away from their familiar home environment during the first 3 weeks of life (Novel) and their matched controls that remained in the home environment (Home) were observed as adults in a 5-min free dyadic social interaction. An analysis of this brief social encounter between two strangers revealed that episodes of mutual social interaction are typically initiated by one of the rats investigating the other member of the pair and terminated by one of the rats ceasing investigation. While Novel and Home rats showed similar frequencies of initiations (p>0.20), the initiations made by Novel rats were reciprocated more frequently than those made by Home rats (p=0.04). Of all the episodes of interaction, more were terminated by Novel rats than by Home rats (p=0.05). Therefore, by being more "successful" in beginning and ending episodes of mutual interaction, Novel rats appear to have a greater control over their social environment than Home rats. These results suggest that brief neonatal stimulation can have long lasting impact on subtle yet socially meaningful aspects of dynamic social interaction. Furthermore, a greater proportion of reciprocation was received by rats with lower basal stress hormone (CORT) concentrations (p=0.04) and by rats with a rightward turning preference (p=0.03), suggesting that HPA axis and patterns of functional brain asymmetry may both be involved in this apparent social control.

Negative maternal behavior predicts offsprings' spatial cognition

A.M. Korzekwa, E. Jaetao, B.C. Reeb, and A.C. Tang
Abstr. ISDP, 2006.

Good maternal care behaviors such as licking and grooming and arched-back nursing have been considered as positive predictors for the development of offspring's stress response system, which in turn modulates cognitive function. Here we present findings that point to the role of negative maternal care behaviors in shaping offspring's cognitive development. We measured two general classes of maternal behavior during PND 1-10: (1) positive maternal care behaviors including licking and grooming of the rat pups and arched-back nursing and (2) negative maternal care behaviors including dropping and burying of pups, stepping on pups, carrying pups in the mouth for an extended period, and disturbing nest by plowing through it. Negative maternal care behaviors were observed immediately after neonatal handling. Positive maternal care behaviors were observed throughout the day following the observation procedure used by Champagne and colleagues (2003). Offspring's spatial memory was measured in the working memory version of the Morris water task at the juvenile stage (P63-68). We found that the more negative maternal care behavior the dam exhibited, the poorer the performance in spatial memory of her offspring (Rs=0.631, p=0.016). In contrast, positive maternal care measures obtained during either the period immediately after handling or the periods throughout the day explained little of the variance in the offspring performance (Rs= -0.229; Rs= -0.104, p>0.20). These results suggest that the presence of bad maternal behavior may be a more powerful predictor for her offspring's cognitive development than the presence of good maternal behavior.

Ability to obtain social interest is impaired by neonatal oxygen deprivation

M. Nakazawa, A.M. Korzekwa, K.G. Akers, and A.C. Tang
Abstr. ISDP, 2006.

Perinatal trauma is known to modify the development of social behavior in both humans and rodents. Here we examined in the rat how neonatal anoxia, a form of perinatal trauma, and subsequent novelty exposure interact to modify an animal's level of interest in social interactions with conspecifics and ability to obtain reciprocal social interest from conspecifics. On postnatal day (P) 1, litters of rat pups were exposed to either 100% N2 gas (Anoxia) or room air (Control) for 25 min. Within each of the Anoxia or Control litters, half of the pups were further individually exposed to relatively novel non-home cages for 3 min daily during P2-21 (Novel) while the other half remained in the home cage (Home). At the onset of adulthood (P100), social interaction was observed among 28 pairs of male rats during one 5-min session. Social interest in a conspecific was indexed by the frequency of attempts to initiate an episode of social interaction (mutual sniffing); the ability to elicit reciprocal interest from a conspecific was indexed by the proportion of initiations that were reciprocated by the conspecific. Preliminary analyses revealed that, although the frequency of initiations did not differ between Control and Anoxia rats (p > 0.20), the proportion of reciprocated initiations was significantly reduced by neonatal anoxia (p < 0.05) and no significant effect of novelty exposure was detected. These results suggest that neonatal trauma induced by oxygen deprivation can impair one's ability to obtain reciprocal social interest.

Reliable detection of bilateral activation in human primary somatosensory cortex by unilateral median nerve stimulation

M.T. Sutherland and A.C. Tang
Abstr. Soc. Neurosci., 2006.

In non-human primates, a bilateral representation of unilaterally presented somatosensory information can be found at the earliest stages of cortical processing as indicated by the presence of neurons with bilateral receptive fields in the hand region of primary somatosensory (SI) cortex. In humans, bilateral activation of SI is considered controversial due to highly variable detection rates for an ipsilateral response (ranging from 3% to 100%) across different studies. In the present study, to improve the detection rate of ipsilateral activations, which are expected to be weak relative to the simultaneous contralateral activations, we applied second-order blind identification (SOBI), a blind source separation algorithm, to high-density EEG data. Previously, SOBI has been shown effective in isolating weak signals from temporally overlapping large amplitude signals. Using established criteria, SOBI components corresponding to focal activation of the left- and right-SI were identified and somatosensory evoked potentials (SEPs) from these two brain regions were obtained. In single subject analysis, statistically significant SEPs in response to unilateral stimulation were detected not only from contralateral-SI but also from ipsilateral-SI. Furthermore, such ipsilateral responses were observed in both the left and right hemispheres of all 10 subjects studied. These results demonstrate that unilateral stimulation of the median nerve, whether applied to the left or right wrist, can activate both the left- and right-SI, raising the possibility that in humans, unilateral sensory input may be bilaterally represented at the earliest stage of cortical processing.

Spatiotemporal Dynamics of SOBI Recovered EEG Components: A Coherence and Granger Causality Analysis

Y. Zhang, A.C. Tang, M.T. Sutherland, and M. Ding
Abstr. Soc. Neurosci., 2006.

Human scalp electroencephalogram (EEG) records a superposition of neural activity from multiple brain sources. As such, it is difficult to assess cortical network dynamics based on sensor level EEG signals. Recent work has shown that second-order blind identification (SOBI) is a technique that can recover the neural source signals from their mixtures. The continuous time series associated with the recovered spatial SOBI components can then be treated as generated by a multivariate stochastic process and subjected to Adaptive MultiVariate AutoRegressive (AMVAR) modeling, yielding power, coherence and Granger causality spectra. Here this approach is applied to study high density EEG (128 channels) activity recorded from eight subjects during five minutes of closed-eyed rest. Our main findings are as follows. First, the power spectra of the component time series, which reflect synchronous oscillatory activities of local neuronal ensembles, show clear and component-specific features: strong alpha oscillations (9-11 Hz) associated with visual components, theta oscillations (5-7 Hz) associated mainly with front midline components, and mu (9-11 Hz) and beta (15-30 Hz) oscillations associated predominantly with sensorimotor components. Second, the coherence analysis reveals the formation of distinct functional networks in the aforementioned frequency bands. The functional connectivity is further dissected by the Granger causality analysis which reveals asymmetrical, sometimes unidirectional, interaction patterns within each pair of interacting components. Third, the SOBI recovered components and the temporal dynamics associated with these components are found to be highly reproducible across subjects. This suggests that the spontaneous ongoing EEG activity represents a highly organized brain state and understanding the spatiotemporal organization of this state may facilitate our understanding of how sensory and motor processing perturbs this state.

Maternal modulation hypothesis as an alternative to maternal mediation hypothesis

B.C. Reeb, M. Sharifi, R.D. Romeo, B.S. McEwen, and A.C. Tang
Abstr. Soc. Neurosci., 2006.

The maternal mediation hypothesis has been recently revived to account for the effects of neonatal stimulation on offspring physiology and behavior (Liu et al., 1997). Recently this hypothesis has been challenged by primate studies that support an alternative hypothesis of stress inoculation (Parker et al., 2006). Here, we present an experiment designed to test a novel hypothesis-the maternal modulation hypothesis. Our hypothesis states that early stimulation has a direct effect on the pups' physiology and behavior and that the magnitude as well as direction of this stimulation effect can be modulated by the dams' post-stimulation behavioral patterns. Specifically, the uncertainty or variability in the post-stimulation maternal behavior serves to shorten or prolong the stimulation effect on the pups' HPA axis, thereby modulating the effects of direct stimulation, hence jointly shaping the offspring's development (Jones & Tang, 1999). To prevent maternal individual differences from confounding the direct pup stimulation effect, we used a split-litter design in which half of each litter was subject to early stimulation while the other half served as controls during the first 3 weeks of life. From a total of 22 litters, 53 male rat pups were exposed to a novel cage for 3 min daily (Novel) while the 53 littermate controls remained in the home cage (Home). Post-novelty exposure variability in maternal licking-grooming negatively correlated with the within-litter novelty effect on offspring's episodic memory measured in the working memory version of the Morris water task [Rs = -.435, p = 0.043]. Furthermore, maternal baseline CORT is negatively correlated with within-litter novelty effect in episodic memory [Rs = -.530, p = 0.013]. Finally, Novel pups showed an overall better episodic memory than the Home pups [F(1,21) = 5.254, p = 0.032]. These results confirm the predictions made by the maternal modulation hypothesis and suggest that neonatal novelty exposure has a separate direct effect on the pups and that maternal caretaking and physiology interact with such effects to determine the effectiveness of such early life manipulation. These findings indicate that the effects of neonatal stimulation are best explained by the maternal modulation hypothesis and cannot be explained by maternal mediation alone.

Maternal modulation hypothesis as an alternative to maternal mediation hypothesis

B.C. Reeb, R.D. Romeo, K.G. Akers, B.S. McEwen, and A.C. Tang
Abstr. Int. Congress Neuroendoc., 2006.

The maternal mediation hypothesis has been recently revived to account for the effects of neonatal stimulation on offspring physiology and behavior (Liu et al., 1997). This hypothesis seems widely taken as implying that neonatal stimulation itself does not produce any effects aside from changing maternal behavior. Here, we present an experiment designed to test an alternative hypothesis-the maternal modulation hypothesis. This hypothesis states that early stimulation has a direct effect on the pups' physiology and behavior and that the dams' behavioral and physiological states subsequent to such stimulation can serve to modulate this direct effect, thus interacting to jointly shape the offspring's development (Jones & Tang, 1999). We used a split-litter design in which half of each litter was subject to early stimulation while the other half served as controls during the first 3 weeks of life. Specifically, from a total of 21 litters, 49 male rat pups were exposed to a novel cage for 3 min daily (Novel) while the 51 littermate controls remained in the home cage (Home). We indexed maternal individual differences in the stress response system by measuring the dams' circulating basal corticosterone concentration 5 days after weaning. We assessed the early stimulation effect in a working memory version of the Morris water task on postnatal day 32-37. ANCOVA with Maternal Basal CORT as the covariate and Novelty as a dummy variable revealed that (1) Novel pups showed better working memory than the Home pups [F(1,19) = 11.156, p = 0.003] and (2) the influence of the maternal individual differences in her HPA function on her pups' working memory differed between the Novel and Home rats (interaction effect between Novelty and Maternal CORT measure [F(1,19) = 5.461; p = 0.031]). These results indicate that neonatal novelty exposure itself resulted in a significant enhancement of working memory, separately from the pups' individual differences in working memory that are accounted for by maternal individual differences. Therefore, the effects of neonatal stimulation are best explained by the maternal modulation hypothesis and cannot be explained by maternal mediation alone.

Characterization of cortical network dynamics: Integrating SOBI and Granger Causality

M.T. Sutherland, Y. Zhang, M. Ding, and A.C. Tang
Abstr. Org. Human Brain Mapping, 2006.

EEG and MEG record a mixture of signals from distributed neural sources. As such it is difficult to assess cortical network dynamics based on raw EEG and MEG data. Second-order blind indentification (SOBI) is a blind source separation algorithm that recovers the source signals from their mixtures by utilizing source temporal coherence. This algorithm has been applied to EEG and MEG data to extract signals that are localized to specific brain regions and has been validated against known noise sources and neuronal sources[1]. When applied to high-density "resting" EEG data (128 channels) obtained without explicit stimulus presentation to the subjects and without explicit task requirements, SOBI was able to extract components with focal spatial origin [2]. Here we explored cortical network dynamics underlying resting EEG activity by applying multivariate time series analysis to SOBI recovered components. The continuous time series associated with the spatially focal components were treated as generated by a multivariate stationary stochastic process. Power, coherence and Granger causality spectra were used to assess the property of this process [3]. We found that (1) the power spectra of these components, which reflect synchronous oscillatory activities of local neuronal ensembles, show clear and distinct patterns of peaks in either the theta, alpha, or beta frequency bands; (2) pairs of these components show distinct patterns of coherence in these frequency bands, indicating that these oscillations bind regional neuronal ensembles into large-scale cortical networks; (3) the causal influences among pairs of components are directional or asymmetrical, with the feedforward influence peaking at one frequency band (e.g. alpha) and the feedback influence peaking at another (e.g. theta). These preliminary findings raise the possibility that top-down versus bottom-up influences among any pairs of brain regions may be obtained non-invasively in humans by combining SOBI and Granger causality analysis. How patterns of such causal influences might vary under different stimulus conditions, mental states, psychiatric conditions, or drug influences remain to be examined.
[1]. Tang, A.C., Sutherland, M.T., and McKinney, C.J. (2005). Validation of SOBI components from high-density EEG. NeuroImage, 25(2): 539-553.
[2]. Sutherland, M.T., and Tang, A.C. (2006). Blind Source Separation can Recover Systematically Distributed Neuronal Sources from "Resting" EEG. Proceedings of the Second International Symposium on Communications, Control, and Signal Processing (ISCCP 2006).
[3]. Ding, M., Chen, Y., and Bressler, S.L. (2006). Granger causality: Basic theory and application to neuroscience. In: Handbook of Time Series Analysis, edited by Matthias Winterhalder, Bjorn Schelter, and Jens Timmer, Wiley, in press.

What's going on in your head while you are resting? Decomposing high-density EEG data using blind source separation.

M.T. Sutherland, and A.C. Tang
Abstr. Org. Human Brain Mapping, 2005.

Objective: Second-order blind identification (SOBI)[1] is a unique blind source separation algorithm that can recover correlated sources (i.e., neuronal activations from different brain regions) from their mixtures (i.e., signals recorded by EEG or MEG sensors). SOBI processing leads to an improved ability to detect neuronal activations that are otherwise difficult or impossible to detect[2] and an enhanced capability to measure single-trial evoked responses[3]. The correspondance between known sources and SOBI components has been validated using noise sourcesand well characterized neuronal sources[4]. One of the most attractive porperties of SOBI is that it exploits the time coherence of source signals to decompose the mixture of activity recorded at the sensors. SOBI uses information inbedded in the temporal structure of the continuous EEG for source separation. As such, detailed temporal characteristics in the ongoing activity from underlying brain sources can provide useful information for source separation. This property of SOBI suggests that the recovery of neuronal sources may be possible without the nedd for stimulus-locked evoked responses. In the present study, we investigated this possibility and attempted to recover sources of brain activity from resting EEG.
Methods: EEG signal from 10 subjects were reorded in an electrically shielded room with a 128-channel EEG system (SynAmps, NeuroScan). During data collection subjects were instructed to sit quietly with eyes closed in a relaxed position. Approximately five minutes of continuous resting EEG was recorded and subsequently used as input to the SOBI algorithm. The 128-channel EEG data were decomposed into 128 recoverd components each of which corresponded to a recovered putative source that contributed to the scalp EEG signal. Each SOBI-recovered putative source had a time course of activation and a sensor space projection that specified the efffect of that component on each of the electrodes.
Results and Discussion: Neuronal sources of interest were identified by the pattern of activation in the component's current source density scalp map and further verified by source localization using an equivalent current dipole model (BESA). Approximately a dozen of the 128 SOBI-components were reliably identified across all subjects, which were analagous to those components recovered from EEG data collected from the same indivivuals during a preceding event-related potential (ERP) experiment. Figure 1 shows examples of four neuronal sources of interest that were reliably identified across all subjects. In addition to cross subject identification, these sources were also repeatedly detected in the same subject when studied during separate recording session. The temporal profile of these sources corresponded to what would be expected given the location of the components, for example, 8-13Hz rythmic activity in the occipitoparietal alpha and sensorimotor mu components.
The implications of SOBI's ability to recover neuronal sources in the absence of ERPs is non-trivial.
References:
1. Belouchrani, A. Aed-Meraim, K., Cardoso, J., and Moulines E. 1997. IEEE Trans. on Signal Proc. , 45, 434-444.
2. Tang, A., Peralmutter, B., Malaszenko, N., Phung, D., and Reeb, B. 2002. Neural Computation, 14, 1827-1858.
3. Tang, A., Peralmutter B., Malaszenko, N. and Phung, D. 2002. NeuroImage,17, 1773-1789.
4. Tang. A., Sutherland, M., and McKinney, C. in press. NeuroImage.

Classifying Single-trial Performance in a Target Detection Task from High Density EEG.

A.C. Tang, K.C. Chang, M.T. Sutherland, C.J. McKinney
Abstr. Org. Human Brain Mapping, 2004.

The goal of the present study was to predict subject responses in a simple target detection task from the EEg. Combining a blind source separation algorithm (BSS) with conventional classification methods, we sought to determine whether a subject made a correct or incorrect response on a given trial using single-trial EEG. The detection task required a subject to respond with a right-handed button press if the middle line in a 5x5 array oflines was tilted (see Fig. 1). The tilt of the line located at the center of the array varied between 0-45 degrees from vertical, by 5 degree increments to create a range of difficulty (less tilt more difficult to detect). Half of the trials were "catch trials" (no tilt) and the other half were equally distributed between 5 and 45 degrees of tilt. Thus, correct responses consisted of making button presses when a tilted line was present and withholding button presses when a titled line was absent. Similarly incorect responses also included both the making and withholding of motor responses. We collected continuous 128-channel EEG throughout the taskperformance and subsequently decomposed the scalp EEG signals into 128 putative underlying sources (components) by appling a BSS algorithm. Previously, using second-order blind identification (SOBI), one such BSS algorithm, we were albe to detectneuronal source activation that was otherwise difficult to detect and to measure single-trial response onset times in a greater number of trials in comparision to measurements obtained directly from the sensors (Tang et. al., Neural Computation; NeuroImage) In the present study, we found that without SOBI pre-processing, a combination of conventional classification methods could correctly classify single-trial correct vs. incorrect responses in ~77% of trials. In contrast, when the scalp EEG data were first processed by SOBI to generate a set of SOBI components, which in turn were used as inputs to the conventional classification methods, single-trial behavioral responses were correctly classified in ~91% of the trials (Fig. 2). When these overall classification results were assessed across multiple levels of difficulty, the classified responses derived from SOBI components (predicted) closely matched the actual observed responses made by the subjects (observed) across the full range of task difficulty (Fig. 3 n=5 subjects). These results demonstrated that correct vs. incorrect behavioral responses in a target detection task can be determined at each single-trial from EEG signals corresponding to that trial. Furthermore, SOBI pre-processing can significantly increase the overall success rate of prediction from below 80% to over 90%.

Neonatal Stimulation and Adult Social Control.

K.G. Akers, M. Nakazawa, B.C. Reeb, A.C. Tang, R.D. Romeo, B.S. McEwen.
Abstr. Soc. Behav. Neuroendocrinol., 2004.

Neonatal novelty exposure, an early life stimulation method, has long lasting effects on learning and memory, synaptic plasticity, and the function of the hypothalamic-pituitary-adrenal (HPA) axis throughout life. We exposed 12 male rat pups to a novel cage for 3 min daily during the first three weeks of life (Novel) while the 12 controls remained in the home cage (Home). At 7-8 month of age, we examined social interactions among Novel-Home pairs during two 5-min social exposures. Eight months later, blood samples were collected at the time of sacrifice. Novel rats showed significantly reduced basal blood CORT concentration than Home rats (t=2.710, p<0.005, df=20). Two social control (SC) scores were derived. SCT score equals the number of episodes of social interaction that was initiated and terminated (T) by the same rat divided by the total number of episodes initiated by that rat. SCD equals the number of episodes of social interaction that was initiated by one rat and was followed by its domination (D) over the other divided by the total number of episodes initiated by that rat. Novel rats increased their social control behaviors while Home rats showed an opposite pattern of change over repeated social interactions (novelty by session interaction SCT: F(1,11)=7.307, p=0.021; SCD: F(1,11)=7.208, p=0.021). These observations suggest that Novel rats exercise more social control by “calling the shots” while Home rats tend to give up such control over repeated social interactions. These findings indicate that very brief and transient neonatal stimulation can lead to profound changes in the capacity for social control and HPA regulation during adulthood.

Modification of Social Memory and Brain Asymmetry by Neonatal Novelty Exposure.

Akaysha C. Tang, Bethany C. Reeb.
Abstr. American Psych. Soc., 2004.

In humans, social recognition from faces appears to involve the left and right brains asymmetrically. We show that, in rodents, early life stimulation that induces brain asymmetry also enhances social recognition memory. Furthermore, individual differences in social memory can be predicted by individual differences in brain asymmetry. Previously, we showed that neonatal novelty exposure, a brief and transient early life stimulation method, can result in long lasting asymmetric changes in synaptic plasticity, both long-term potentiation (LTP) and depression (LTD), hippocampal volume, and “handedness”. Parallel to this early stimulation-induced functional lateralization, we found that neonatal novelty exposure enhances memory function as demonstrated by enhanced spatial working memory in the Morris water maze task and enhanced retention in odor discrimination. Many have speculated possible functional significances of brain asymmetry. In this study, we evaluated a possible link between social recognition and brain asymmetry and its modulation by early life stimulation. We divided each of the 9 new born litters of rats into two groups, one exposed to a new cage for 3 min daily for the first three weeks of life (NOVEL) while the other remained in the home cage (HOME). During adulthood (7 months of age), NOVEL and HOME rats were exposed to each other in pairs on two consecutive days in four 5 min sessions: Day 1: S1-S3, with 10 and 2 min inter-trial interval; Day 2: S1. Short-term social recognition is indexed by a short-term habituation score, defined as (S1-S2)/S1 and long-term recognition by a long-term habituation score defined as (D1S1 – D2S1)/D1S1. Prior to social exposure on each day, rats were first exposed to the testing cage for 5 min individually for cage habituation. During this initial exposure to the testing environment, asymmetric brain activation was indexed by a lateralization score defined as (R-L)/(R+L), where L and R indicates the frequency of left and right turns. We found both short- and long term habituation were significantly greater in the NOVEL than HOME rats (short-term: t=2.549; p<0.01; df=21; long-term: t=2.259; p<0.025; df=21). This finding indicates that neonatal novelty exposure can produce long lasting enhancement in social recognition memory that persists into adulthood. Furthermore, we found that only NOVEL rats showed 24 hr habituation to a previously encountered conspecific (t =2.791; p < 0.01; df =11), extending the duration of social recognition memory from the previously reported < 2 hr to at least 24 hrs. Parallel to changes in social memory, NOVEL rats showed significantly greater right turn preference than the HOME rats (t=1.860; p<0.05; df=21), consistent with our previous reports of early stimulation-induced right-shift. Individuals with a greater right turn preference showed greater short-term habituation (rs=0.521, p<0.05) even after the neonatal novelty effect was partialed out ( partial r=0.445, p<0.019; n=23). This correlation provides support for the hypothesis that brain asymmetry may be an important source of influence on social recognition memory.

Blind Source Separation Allows for the Reliable Detection of Ipsilateral Primary Somatosensory Cortex Activation.

M.T. Sutherland, C.J. McKinney, A.C. Tang.
Abstr. Soc. Cog. Neurosci. 2004.

Median nerve stimulation activates primary somatosensory (SI) cortex contralateral to the side of stimulation. Less obvious is the activation of ipsilateral SI. While some positive results have been reported using subdural (e.g., Noachtar, 1997) and MEG (e.g., Korvenoja et al., 1995; Kanno et al., 2003) recording techniques, only in a small subset of subjects could ipsilateral activation be detected in previous median nerve stimulation experiments. A possible hypothesis accounting for such variability is that ipsilateral responses are generated in most people but could not be detected using conventional approaches because other stronger sources of activation can “mask” the weaker ipsilateral responses. To test this hypothesis we applied Second-Order Blind Identification (SOBI), a blind source separation (BSS) algorithm to the analysis of 128 channel continuous EEG data. Using SOBI we isolated and identified components whose time course of activation and spatial localization corresponded to those of SI, along with other neuronal sources and typical artifacts. SOBI components that corresponded to SI displayed somatosensory evoked potentials (SEPs) to both contra- and ipsilateral stimulation. Although ipsilateral SEPs varied between subjects, both left and right SI responded to ipsilateral stimulation in all subjects studied thus far. These results suggest that SOBI is an effective signal processing tool for revealing aspects of brain function that were otherwise difficult to detect using conventional approaches.

Source Localization from High Density EEG Data During a Real World Task.

A.C. Tang, C.J. McKinney, M.T. Sutherland, L. Parra, B.C. Reeb, N.A. Malaszenko, A. Gerson, P. Sajda.
Abstr. Soc. Neurosci. 2003.

Typically EEG data are collected under laboratory conditions, which depart drastically from any real world operational settings. Repetitions of hundreds of trials and control of eye movements have been necessary for obtaining useful EEG data. In a typical visual task, subjects are often asked to fixate on a stationary point to minimize ocular artifacts in the EEG data that are associated with eye movements. In the real world, such tight control is not possible. Therefore, identifying the underlying neuronal sources without these experimental constraints represent a major technical challenge. Previously, we were able to recover neuronal sources of activation that could not be recovered using conventional source modeling methods by applying second-order blind identification (SOBI), an independent component analysis algorithm to continuous MEG data (Tang et al., 2002). Here, we extended this method of signal processing from MEG to EEG data collected during the Warship Commander Task (WCT)a realistic ship based navy command and control environment that contains multiple tasks and enlists spatial, verbal, and decision making processes (St. John et al., 2002). We were able to reliably identify three common types of neuronal sources of activation across all subjects and task scenarios: (1) a frontal-ocular activation which appeared to be associated with task-related eye movement, (2) an occipital-parietal activation which is indicative of typical early visual processing, and (3) a synchronized anterior-posterior activation which is likely to reflect the activation of an attentional network.

SOBI Analysis of Electroencephalographic Data and Exploration of Critical Parameters.

C.J. McKinney, M.T. Sutherland, N.A. Malaszenko, B.C. Reeb, A.C. Tang.
Abstr. Soc. Neurosci. 2003.

Previously we applied second-order blind identification (SOBI), an independent component analysis (ICA) algorithm to the analysis of magnetoencephalographic (MEG) data. Using SOBI as a preprocessing tool, neuronal source activation that was otherwise difficult to detect can be detected with increased success rate in comparison to using conventional data analysis (Tang et al., Neural Comp. 2002). Furthermore, single trial response onset times can be measured in a greater number of trials with reduced false detection rate in comparison to measurement obtained directly from the best sensors (Tang et al, NeuroImage 2002). Here, we extended SOBI to the analysis of high density (128 channel) EEG data collected during mixed unilateral and bilateral visual and somatosensory stimulation (150 trials per condition), to specifically test SOBI's ability in identifying temporally overlapping neuronal source activations. We found that SOBI is able to recover components that correspond to activation of the expected left and right visual and somatosensory sources. Temporally, these components displayed characteristics sensory responses appropriate for the corresponding modality and have a significantly high signal to noise ratio than the signals from the best sensors. Spatially, these components can be modeled by equivalent current dipoles located in the appropriate sensory cortices. Simultaneously, SOBI also identifies and localizes components that correspond to known bad channels, thus providing direct validation for the SOBI source separation process. These results show that SOBI is suitable for the analysis of functional brain imaging data obtained from both MEG and EEG. Finally, we will also explore the sensitivity of the SOBI results to a variety of parameters used by the SOBI algorithm.

Detection and Qualitative Analysis of Single-Trial Brain Evoked Response Combining Multi-Scale Metrics with Independent Component Analysis.

B.C. Reeb, T.A. Loring, D.E. Worth, N.A. Malaszenko, A.C. Tang.
Abstr. Soc. Neurosci. 2003.

With second-order blind identification (SOBI), an ICA algorithm, one can separate both environmental noise and major artifacts from signals associated with brain activity. Via low-pass filtering and thresholding, we can further detect the onsets of evoked response potentials (ERPs) in a high percentage of individual trials without excessive false detection (Tang et al, Neuroimage 2002). However, other methods of detection should be less susceptible to noise associated with ongoing background activity in the MEG or EEG data. We have implemented a flexible, multiscale metric that quantifies how any given point in a time-series differs from the point of response onset, or response peak in a canonical response. The multiscale filtering behind the metric is similar to the filtering in the continuous wavelet transform, as has been used for onset detection (Mekle et al, Proc. SPIE vol. 4119, 2000). The precise filter banks, and the resulting metrics, are customized using visualization techniques and have improvements over the continuous wavelet transform, suggested by the steerable multiresolution pyramids used in image processing (Simoncelli et al, IEEE Trans. Information Theory, 1992). We will apply these methods to SOBI components from real and simulated MEG or EEG data for detecting single-trial ERPs and compare the relative performance and accuracy of our metrics against both the threshold with low-pass and the multiscale singularity detection.

Evidence for Plasticity in Somatosensory Cortex After Paired Right and Left Median Nerve Stimulation.

M.T. Sutherland, C.J. McKinney, A.C. Tang.
Abstr. Soc. Neurosci. 2003.

We have taken the idea of associativity, typically discussed at the microscopic level as long-term potentiation (LTP) of synaptic strength measured at the tip of an electrode (< 10 x 10 micron), and attempted to generalize it to a macroscopic level as persistent change of synaptic strength measured non-invasively from a large piece of cortex (10 x 10 mm). Can the simultaneous inputs from two neural pathways converging onto a large population of neurons cause a change in the evoked response of that population when stimulation from a single pathway is later applied? Previously, Braun et al. (2000) have shown that massed practice is able to alter the organization of single finger representations in somatosensory cortex. Here we attempt to induce cortical reorganization between the right and left primary somatosensory cortices (SI) using median nerve stimulation. Since SI can be activated by both contra- and ipsilateral inputs (Korvenoja et al., 1995; Schnitzler et al., 1995; Kanno et al., 2003), we manipulated the delays between contra- and ipsilateral median nerve stimulations to provide convergent inputs to SI. Applying independent component analysis (ICA) to EEG data, we isolated and identified components whose time course of activation and spatial localization corresponded to those of SI. Preliminary data showed that co-activation of the contra- and ipsilateral pathways in a single recording session can produce reorganization between left and right SI. This plasticity induced by co-activation is easily initiated and long-lasting. Several months after the initial paired stimulations, ipsilateral SI showed an increased activation that is synchronized to contralateral SI activation.

Effects of Estrogen Replacement on Hypothalamic-Pituitary-Adrenal (HPA) Axis, Social Recognition, and Open Field Behavior.

M. Nakazawa, A.C. Tang, B.C. Reeb, H.M. Sisti, R.D. Romeo, B.S. McEwen.
Abstr. Soc. Neurosci. 2003.

We investigated the effect of estrogen replacement (ER) on HPA axis response to stress associated with a behavioral task and on social recognition memory and open field behavior. Twenty-three 10 weeks old female mice were ovariectomized and implanted with the following: 0.00 (OVX), 0.18 (OVX-LOW-E), or 0.72 (OVX-HIGH-E) mg of 17ß estradiol. Three weeks after surgery, activity level and the latency to leave the center were measured in an open field for a single trial of 60 sec. Four weeks later, social recognition was measured by habituation of social investigation after repeated 5-min dyadic social exposures on two consecutive days. Two days later, blood samples were collected to measure corticosterone stress response 90 min after the onset of four swimming trials similar to those in a Morris water task. The implants were effective as indicated by the higher normalized uterus weight observed among the mice with higher level of estrogen replacement (p< 0.001). ER induced a change in the HPA axis as indicated by a decreasing trend in the blood corticosterone concentration across the OVX, OVX-LOW-E, and OVX-HIGH-E groups (p < 0.05). This result is consistent with the hypothesis that ER increases feedback sensitivity of the HPA axis to stress induced by a commonly used spatially learning task. Twenty four hour habituation in social investigation was found among the HIGH-E but not the other groups ( p < 0.01), indicating an enhanced social recognition by estrogen. The mice receiving higher levels of estrogen also took longer to leave the center of the open field and traveled less during the 1 min interval (p < 0.001, p < 005, respectively). These results raise the possibility that the effect of long-term ER on learning memory may be mediated by a modification of the HPA axis.

Neonatal Novelty Exposure Modulates Sex Differences in Turning Bias.

K.G. Akers, B.C. Reeb, A.C. Tang.
Abstr. Soc. Neurosci. 2003.

Neonatal novelty exposure has been shown to induce a right-hemisphere dominance in adult rats, as evidenced by a right-shift in hippocampal volume (Verstynen et al 2001), a left-shift in paw preference (Tang & Verstynen 2002), an enhancement of synaptic plasticity in the right hippocampus (Zou & Tang 2001,2002), and a right-shift in turning behavior (Tang & Reeb 2003). These changes are in parallel with neonatal novelty induced enhancement of learning and memory performance in a variety of tasks (Tang 2001; Reeb et al 2001), adding to the idea that brain asymmetry may underlie cognitive function. We investigated how neonatal novelty exposure affects sex differences in cerebral lateralization by examining a turning bias at 7 weeks and 7 months of age. During the first 3 weeks of life, half of a litter was exposed to a novel environment (Novel) for 3 min daily while the other half remained in the home cage (Home). At 7 weeks of age turning behavior was investigated in rats during a 5 min exploration session in a novel cage on two consecutive days. Testing was repeated at 7 months of age. We found a significant interaction effect between sex and neonatal novelty exposure on turning bias across days (p<.01). In the Novel animals, females had a significantly greater left-turn bias than males (p<.025), while the opposite pattern was found in the Home animals (p<.025). These results suggest that neonatal novelty exposure modulates sex differences in cerebral lateralization. A similar sex by neonatal novelty exposure interaction has also been found in open field disinhibition (Tang et al 2003) and social recognition memory (Reeb et al 2002).

Daily Changes in the Rat's Spontaneous Turning Behavior During the First Three Weeks of Post-Natal Life.

M. Nakazawa, B. Reeb, A. Murley, A.C. Tang.
Abstr. ISDP. 2003.

Exposing rat pups to a novel environment for the first three weeks of life has been shown to induce, during adulthood, an increase in right brain dominance measured by changes in hippocampal volume (Verstynen et al 2001), synaptic plasticity (Tang 2003, The Asymmetrical Brain, MIT Press), and a corresponding increase in left dominance in peripheral paw preference (Tang and Verstynen 2002). In this study, we made daily observations on 9 litters of pups during the first three weeks of life in an attempt to describe the initial behavioral indicators of this rightwards shift in brain asymmetry. Pups were observed for 3 min daily in a small housing cage lined with fresh saw dust. Their movements were video taped and analyzed off line for left and right turns. We found a significant linear trend from week 1 to 3 (F(1,8)=6.21; p=0.037), indicating a rightwards shift towards symmetry, occurring during the first 3 weeks of life. When each week was examined separately, we found a significant left turn preference during the first post-natal week (t=4.340; p<0.01; df=8) and no turning preferences during week 2 and 3 (p>0.20). Thus, major changes in turning preference occurred within the first week of life.

Sex Differences in Social Recognition Memory in a Longitudinal Study of the Rat.

B.C. Reeb, A.C. Tang.
Abstr. ISDP. 2003.

Among adult rats, social recognition memory in females was considered longer lasting than males and neither lasting beyond 2 hours (Bluthe & Dantzer, 1990). No sex difference in social recognition was observed during adolescence (Thor & Holloway, 1982). To replicate our previous finding that neonatal stimulation can prolong social recognition memory to at least 24 hrs among the male rats (Tang et al 2003) and to evaluate potential sex differences at both adolescence and adulthood, we conducted a longitudinal study in which 30 male and 40 female pups were first exposed to a neonatal novelty procedure during the first 3 weeks of life and then tested for their social recognition memory at 7 weeks (adolescence) and 7 months of age (adulthood). We measured short- and long-term habituation (STH and LTH) to a previously encountered conspecific. Contrary to the previously reported sex difference in adults, we found that males showed a greater STH than females (p< 0.05), suggesting a longer lasting short-term social memory in males. Furthermore, this sex difference in short-term social memory was also found during adolescence (p<0.05). For long-term (24 hr) social memory, males showed a greater LTH than females during adulthood (p<0.05) and a similar but marginally significant difference during adolescence (p=0.078). Among adult males, the LTH was significantly greater than zero (p<0.05) indicating a social recognition that lasted for at least 24 hrs (replicating our previous findings). These findings suggest that male and female rats differ in both short- and long-term social recognition memory during both adolescence and adulthood.

Neonatal Novelty Exposure Modulates Sex Differences in Turning Bias.

K.G. Akers, B.C. Reeb, A.C. Tang.
Abstr. ISDP. 2003.

Neonatal novelty exposure has been shown to induce a right-hemisphere dominance in adult rats, as evidenced by a right-shift in hippocampal volume (Verstynen et al 2001), a left-shift in paw preference (Tang & Verstynen 2002), an enhancement of synaptic plasticity in the right hippocampus (Zou & Tang 2001, 2002), and a right-shift in turning behavior (Tang & Reeb 2003). These changes are in parallel with neonatal novelty induced enhancement of learning and memory performance in a variety of tasks (Tang 2001; Reeb et al 2001), adding to the idea that brain asymmetry may underlie cognitive function. We investigated how neonatal novelty exposure affects sex differences in cerebral lateralization by examining a turning bias at 7 weeks and 7 months of age. During the first 3 weeks of life, half of a litter was exposed to a novel environment (Novel) for 3 min daily while the other half remained in the home cage (Home). At 7 weeks of age turning behavior was investigated in rats during a 5 min exploration session in a novel cage on two consecutive days. Testing was repeated at 7 months of age. We found a significant interaction effect between sex and neonatal novelty exposure on turning bias across days (p<.01). In the Novel animals, females had a significantly greater left-turn bias than males (p<.025), while the opposite pattern was found in the Home animals (p<.025). These results suggest that neonatal novelty exposure modulates sex differences in cerebral lateralization. A similar sex by neonatal novelty exposure interaction has also been found in open field disinhibition (Tang et al 2003) and social recognition memory (Reeb et al 2002).

Neonatal Novelty Exposure Modulates Adult Corticosterone Stress Response Elicited by the Morris Water Maze Test.

A.C. Tang, B. Reeb, M. Nakazawa, R.D. Romeo, K.G. Akers, B.S. McEwen.
Abstr. ISDP. 2003.

Exposing rat pups to a novel environment for the first three weeks of life has been shown to enhance learning in the working memory version of the Morris water maze task (Tang, 2001). This study examined differences between neonatal novelty exposed (Novel: n=16) and control (Home: n=13) rats in their corticosterone (CORT) response elicited by the water maze task during adulthood (17 months old). Two baseline blood samples were collected two days prior to the water maze. After reaching asymptote in the water maze, rats were tested on two days for their one-trial learning, measured by trial 2 latency, followed immediately by blood collection. The two days of testing differed by the introduction of a surprise exposure to an open field between trials 1 and 2. We found (1) a significant interaction of swim by neonatal novelty treatment effect on post-swim CORT, with the Novel rats showing less swim-elicited increase in CORT than the Home rats (F(1,7) = 12.059; p< 0.01); (2) a marginally significant neonatal novelty effect on the post-swim CORT ( F(1,7)=3.552, p=0.0505 directional), with the Novel rats showing lower serum CORT concentration than the Home; and (3) a significant open field surprise effect on both post-swim CORT (F(1,7)=6.561, p=0.037) and one-trial learning (F(1,7)=6.982, p=.033). These results suggest that the CORT stress response elicited by a familiar water maze test can be modulated by temporally remote early life stimulation and that the interference effect of an unexpected open field exposure on one trial learning in the water maze may be mediated by a parallel interference effect on the circulating CORT concentration.

Application of an ICA Algorithm for the Analysis of Electroencephalographic Data.

A.C. Tang, C.J. McKinney, M.T. Sutherland.
Abstr. Soc. Cog. Neurosci. 2003.

Previously we applied second order blind identification (SOBI), an independent component analysis (ICA) algorithm to the analysis of magnetoencephalographic (MEG) data. Using SOBI as a preprocessing tool, neuronal source activation that was otherwise difficult to detect can be detected with increased success rate in comparison to using conventional data analysis (Tang et al., Neural Comp. 2002). Furthermore, single trial response onset times can be measured in a greater number of trials with reduced false detection rate in comparison to measurement obtained directly from the best sensors (Tang et al, Neuroimage 2002). Here, we extended SOBI to the analysis of high density (128 channel) EEG data collected during mixed unilateral and bilateral stimulation across three major sensory modalities (150 trials per condition), to specifically test SOBI's ability in identifying temporally overlapping neuronal source activations. We found that SOBI is able to recover components that correspond to activation of the expected left and right visual, auditory, and somatosensory sources. Temporally, these components displayed characteristics sensory responses appropriate for the corresponding modality. Spatially, these components can be modeled by equivalent current dipoles located in the appropriate sensory cortices. Simultaneously, SOBI also identifies and localizes components that correspond to known bad channels, thus providing direct validation for the SOBI source separation process. These results show that SOBI is suitable for the analysis of functional brain imaging data obtained from both MEG and EEG.

Somatosensory to Visual Cross-Modal Interaction: Evidence from Visual Alpha Resetting by Median Nerve Stimulation.

C.J. McKinney, M.T. Sutherland, A.C. Tang.
Abstr. Soc. Cog. Neurosci. 2003.

Cross-modal experiments showed that early sensory processing across modalities are more integrated than previously thought. During audio-visual pairing, visual event-related potentials (VEPs) recorded at the scalp over occipito-parietal cortex were more than the algebraic sum, suggesting a possible modification of neuronal responses within early visual processing areas (Foxe et al 99). Similar results were also obtained from an MEG study in which our group found that the time course of visual activation in early processing areas were sensitive to the visual-audio time-delays and that single-trial VEP onset times of these regions were modified by the paired auditory stimulation. In the present study, applying ICA (see also abs by Tang et al) to EEG data collected during separate periods of visual and somatosensory stimulation. Two separate types of components with prominent alpha band activity were found with ECDs in the occipital or occipito-parietal cortex: one with and one without characteristic VEPs. While leading to characteristic activation of SI, somatosensory (median nerve) stimulation resets the phase of the alpha band oscillation only for the alpha components that did not respond to visual stimulation (components with no VEPs). These results suggest that occipito-parietal synchronized neuronal activations within the alpha band may belong to multiple functionally distinct units that cannot be simply distinguished by their spatial distribution across the scalp and that a subset of these functional alpha units are under tonic modulation by sensory inputs from another sensory modality.

Rapid Interhemispherical Associative Learning in Human Somatosensory Event-Related Potentials.

M.T. Sutherland, C.J. McKinney, A.C. Tang.
Abstr. Soc. Cog. Neurosci. 2003.

Hebb postulated that simultaneous activation of neuronal population leads to increased synaptic connections among the synchronously activated neurons. To test this hypothesis, we delivered left and right somatosensory stimulation (mild median nerve shocks) with multiple delays and varying temporal overlap. Applying second order blind identification (SOBI), an independent component analysis (ICA) algorithm to 128 channel high density continuous EEG data, we isolated and identified components whose time course of activation and spatial localization correspond to those of SI (see also abs by Tang et al), along with other neuronal sources and typical artifacts. somatosensory evoked potentials (SEPs) of left and right SI from the early and late trials during each 150 trial blocks were compared for each time delay to reveal delay-dependent changes. Preliminary result showed that co-activation of the SIs by ipsi- and contra-lateral stimulation can lead to rapid changes in the SEPs in less than 3 min. of stimulation. The magnitude and temporal characteristic of the changes in the SEPs appeared to be modulated by the delay and overlap between paired stimulation. Because of individual differences in inter-hemisphere transfer time, the optimal delays for the maximum learning-induced changes appear to vary between subjects. These preliminary results offer a direct demonstrate of changes in one processing pathway modified by simultaneous inputs from a second pathway. Along with conditioning-induced increase in gamma band coherence (Milner et al 1999), these results offer support for the operation of a Hebbian-like learning mechanism.

Neonatal Novelty Exposure Induces Long Lasting Enhancement in Long Term Depression (LTD) in the Right Hippocampus of Adult Rats.

B.D. Zou, B. Reeb, & A.C. Tang
Abstr. Soc. Neurosci. 2002.

Brief neonatal novelty exposure results in long-lasting enhancement in hippocampal-dependent spatial and social memory (Tang2001; Caplan et al 2001; Reeb et al 2001). Parallel to these learning enhancement, neonatal novelty exposure also enhances long-term potentiation (LTP) in CA1 of hippocampus in adult rats and LTP's sensitivity to the stress hormone corticosterone (Zou etl al 2001; Tang and Zou 2002). In this experiment, we examined long-term depression (LTD) in the rat CA1 of the hippocampus during adulthood. Pups were exposed daily to a non-home environment briefly for 3 min during the first 3 weeks of life (Novel) while the control pups remained at home (Home). Field excitatory postsynaptic potential (fEPSP) was recorded in the CA1 stratum radiatum of hippocampus in 7-8 month old rats. We found that in comparison to Home rats, low frequency stimulation (3Hz,10min) induced greater depression among the Novel rats in the right (but not the left) side of the hippocampus (p<0.05). These results demonstrate that neonatal novelty exposure results in long-lasting (> 6 months) enhancement in LTD in adult rats. Along with