University of Southern California
University of Southern California
Mrs. T.H. Chan Division of Occupational Science and Occupational Therapy
Mrs. T.H. Chan Division of Occupational Science and Occupational Therapy
Twitter Facebook Instagram LinkedIn YouTube

Barbara Thompson PhD

Barbara Thompson PhD

Assistant Professor, joint appointment with the Keck School of Medicine of USC, Department of Pediatrics

Room: CHP 222H
(323) 442-2808
.(JavaScript must be enabled to view this email address)


Barbara Thompson was most recently a research assistant professor at the Keck School of Medicine of USC, Department of Cell and Neurobiology. She had previously been a postdoctoral fellow at Vanderbilt University Medical Center’s Department of Pharmacology. Dr. Thompson has conducted groundbreaking research in the neurobiology of developmental disorders and is contributing to the development of translational science in areas such as autism by developing and testing conceptual models which link clinical problems in human populations to animal studies in the neurosciences. Dr. Thompson’s particular skill in interdisciplinary research collaborations strengthens the USC Chan Division’s critical mass in autism and sensory integration with a special focus on neuroscience research.


Postdoctoral Fellow
2009 | Vanderbilt University Medical Center

Doctor of Philosophy (PhD) in Psychology
2003 | University of Delaware

Bachelor of Science (BS) in Psychology
1997 | Florida State University


Journal Articles

Hiller, L. T., Takata, S., & Thompson, B. L. (2015). Conditioned place preference successfully established in typically developing children. Frontiers in Behavioral Neuroscience, 9, 187. Show abstractHide abstract

Affective processing, known to influence attention, motivation, and emotional regulation is poorly understood in young children, especially for those with neurodevelopmental disorders characterized by language impairments. Here we faithfully adapt a well-established animal paradigm used for affective processing, conditioned place preference (CPP) for use in typically developing children between the ages of 30-55 months. Children displayed a CPP, with an average 2.4 fold increase in time spent in the preferred room. Importantly, associative learning as assessed with CPP was not correlated with scores on the Mullen Scales of Early Learning (MSEL), indicating that CPP can be used with children with a wide range of cognitive skills.

Wilkinson, B., Grepo, N., Thompson, B. L., Kim, J., Wang, K., Evgrafov, O. V., Lu, W., Knowles, J. A., & Campbell, D. B. (2015). The autism-associated gene chromodomain helicase DNA-binding protein 8 (CHD8) regulates noncoding RNAs and autism-related genes. Translational Psychiatry, 5, e568. Show abstractHide abstract

Chromodomain helicase DNA-binding protein 8 (CHD8) was identified as a leading autism spectrum disorder (ASD) candidate gene by whole-exome sequencing and subsequent targeted-sequencing studies. De novo loss-of-function mutations were identified in 12 individuals with ASD and zero controls, accounting for a highly significant association. Small interfering RNA-mediated knockdown of CHD8 in human neural progenitor cells followed by RNA sequencing revealed that CHD8 insufficiency results in altered expression of 1715 genes, including both protein-coding and noncoding RNAs. Among the 10 most changed transcripts, 4 (40%) were noncoding RNAs. The transcriptional changes among protein-coding genes involved a highly interconnected network of genes that are enriched in neuronal development and in previously identified ASD candidate genes. These results suggest that CHD8 insufficiency may be a central hub in neuronal development and ASD risk.

Eagleson, K. L., Campbell, D. B., Thompson, B. L., Bergman, M. Y., & Levitt, P. R. (2011). The autism risk genes MET and PLAUR differentially impact cortical development. Autism Research, 4(1), 68-83. Show abstractHide abstract

Candidate risk genes for autism spectrum disorder (ASD) have been identified, but the challenge of determining their contribution to pathogenesis remains. We previously identified two ASD risk genes encoding the receptor tyrosine kinase MET and the urokinase plasminogen activator receptor (PLAUR), which is thought to modulate availability of the MET ligand. We also reported a role for Met signaling in cortical interneuron development in vitro and a reduction of these neurons in uPAR (mouse ortholog of PLAUR) null mice, suggesting that disruption of either gene impacts cortical development similarly. Here, we modify this conclusion, reporting that interneuron numbers are unchanged in the neocortex of Metfx/fx / Dlx5/6cre mice, in which Met is ablated from cells arising from the ventral telencephalon (VTel). Consistent with this, Met transcript is not detected in the VTel during interneuron genesis and migration; furthermore, during the postnatal period of interneuron maturation, Met is co-expressed in glutamatergic projection neurons, but not interneurons. Low levels of Met protein are expressed in the VTel at E12.5 and E14.5, likely reflecting the arrival of Met containing corticofugal axons. Met expression, however, is induced in E12.5 VTel cells after 2 days in vitro, perhaps underlying discrepancies between observations in vitro and in Metfx/fx / Dlx5/6cre mice. We suggest that, in vivo, Met impacts the development of cortical projection neurons, whereas uPAR influences interneuron maturation. An altered balance between excitation and inhibition has been postulated as a biological mechanism for ASD; this imbalance could arise from different risk genes differentially affecting either or both elements.

Thompson, B. L., & Levitt, P. R. (2010). The clinical-basic interface in defining pathogenesis in disorders of neurodevelopmental origin. Neuron, 67(5), 702-712. Show abstractHide abstract

Human cognitive and social-emotional behaviors are heterogeneous, underscoring the challenges in modeling pathogenesis in disorders of neurodevelopmental origin in which these domains are dysfunctional. In general, animal models for these disorders are built to emulate our understanding of the clinical diagnosis, with mixed results. We suggest the utility of model systems lies in the use of different strategies to perturb hierarchical circuit development, to examine the behavioral dimensions that are most impacted, and to discern the capacity for, and heterogeneity of, neuroadaptation that will then inform treatment strategies.

Thompson, B. L., Stanwood, G. D., & Levitt, P. R. (2010). Specificity of prenatal cocaine exposure effects on cortical interneurons is independent from dopamine D1 receptor co-localization. Journal of Chemical Neuroanatomy, 39(4), 228-234. Show abstractHide abstract

Gestational cocaine exposure in a rabbit model leads to a persistent increase in parvalbumin immunoreactive cells and processes, reduces dopamine D1 receptor coupling to Gsalpha by means of improper trafficking of the receptor, changes pyramidal neuron morphology, and disrupts cognitive function. Here, experiments investigated whether changes in parvalbumin neurons were specific, or extended to other subpopulations of interneurons. Additionally, we examined dopamine D1 receptor expression patterns and its overlap with specific interneuron populations in the rabbit prefrontal cortex as a possible correlate for alterations in interneuron development following prenatal cocaine exposure. Analysis of calbindin and calretinin interneuron subtypes revealed that they did not exhibit any differences in cell number or process development. Thus, specific consequences of prenatal cocaine in the rabbit appear to be limited to parvalbumin-positive interneurons. Dopamine D1 receptor expression did not correlate with the selective effects of cocaine exposure, however, as both parvalbumin and calbindin cell types expressed the receptor. The findings suggest that additional, unique properties of parvalbumin neurons contribute to their developmental sensitivity to in utero cocaine exposure.

Thompson, B. L., & Levitt, P. R. (2010). Now you see it, now you don’t—Closing in on allostasis and developmental basis of psychiatric disorders. Neuron, 65(4), 437-439. Show abstractHide abstract

The mode through which early insults in brain development result in the onset of psychiatric disorders years after impact become a little less mysterious with the report by Niwa et al. that a transient reduction of the schizophrenia risk gene DISC1 can alter prefrontal cortex neurochemistry, architecture, and function.

Thompson, B. L., Levitt, P., & Stanwood, G. D. (2009). Prenatal exposure to drugs: effects on brain development and implications for policy and education. Nature Reviews: Neuroscience, 10, 303-312. Show abstractHide abstract

The effects of prenatal exposure to drugs on brain development are complex and are modulated by the timing, dose and route of drug exposure. It is difficult to assess these effects in clinical cohorts as these are beset with problems such as multiple exposures and difficulties in documenting use patterns. This can lead to misinterpretation of research findings by the general public, the media and policy makers, who may mistakenly assume that the legal status of a drug correlates with its biological impact on fetal brain development and long-term clinical outcomes. It is important to close the gap between what science tells us about the impact of prenatal drug exposure on the fetus and the mother and what we do programmatically with regard to at-risk populations.

Thompson, B. L., & Stanwood, G. D. (2009). Pleiotropic effects of neurotransmission during development: modulators of modularity. Journal of Autism and Developmental Disorders, 39(2), 260-268. Show abstractHide abstract

The formation and function of the mammalian cerebral cortex relies on the complex interplay of a variety of genetic and environmental factors through protracted periods of gestational and postnatal development. Biogenic amine systems are important neuromodulators, both in the adult nervous system, and during critical epochs of brain development. Abnormalities in developmental programming likely contribute to developmental delays and multiple neurological and psychiatric disorders, often with symptom onset much later than the actual induction of pathology. We review several genetic and pharmacological models of dopamine, norepinephrine and serotonin modulation during development, each of which produces permanent changes in cerebral cortical structure and function. These models clearly illustrate the ability of these neurotransmitters to function beyond their classic roles and show their involvement in the development and modulation of fine brain circuitry that is sensitive to numerous effectors. Furthermore, these studies demonstrate the need to consider not only gene by environment interactions, but also gene by environment by developmental time interactions.

D'Arceuil, H., Liu, C., Levitt, P., Thompson, B. L., Kosofsky, B., & de Crespigny, A. (2008). Three-dimensional high-resolution diffusion tensor imaging and tractography of the developing rabbit brain. Developmental Neuroscience, 30(4), 262-275. Show abstractHide abstract

Diffusion tensor imaging (DTI) is sensitive to structural ordering in brain tissue particularly in the white matter tracts. Diffusion anisotropy changes with disease and also with neural development. We used high-resolution DTI of fixed rabbit brains to study developmental changes in regional diffusion anisotropy and white matter fiber tract development. Imaging was performed on a 4.7-tesla Bruker Biospec Avance scanner using custom-built solenoid coils and DTI was performed at various postnatal ages. Trace apparent diffusion coefficient, fractional diffusion anisotropy maps and fiber tracts were generated and compared across the ages. The brain was highly anisotropic at birth and white matter anisotropy increased with age. Regional DTI tractography of the internal capsule showed refinement in regional tract architecture with maturation. Interestingly, brains with congenital deficiencies of the callosal commissure showed selectively strikingly different fiber architecture compared to age-matched brains. There was also some evidence of subcortical to cortical fiber connectivity. DTI tractography of the anterior and posterior limbs of the internal capsule showed reproducibly coherent fiber tracts corresponding to known corticospinal and corticobulbar tract anatomy. There was some minor interanimal tract variability, but there was remarkable similarity between the tracts in all animals. Therefore, ex vivo DTI tractography is a potentially powerful tool for neuroscience investigations and may also reveal effects (such as fiber tract pruning during development) which may be important targets for in vivo human studies.

Parlaman, J. P., Thompson, B. L., Levitt, P., & Stanwood, G. D. (2007). Pharmacokinetic profile of cocaine following intravenous administration in the female rabbit. European Journal of Pharmacology, 563(1-3), 124-129. Show abstractHide abstract

Prenatal cocaine exposure in a rabbit intravenous model has revealed selective disruption of brain development and pharmacological responsiveness. We therefore examined the pharmacokinetic properties of cocaine in this model. Dutch-belted rabbits were surgically implanted with a catheter in the carotid artery, allowed to recover, and then injected intravenously with a cocaine bolus. Cocaine and benzoylecgonine concentrations were measured in arterial blood plasma and analyzed by nonlinear regression and noncompartmental analyses. Peak cocaine concentration occurred by 30s, was transient, and distribution was rapid. The profile of cocaine in the rabbit is similar to that observed in humans using cocaine at recreational doses.

Thompson, B. L., & Rosen, J. B. (2006). Immediate-early gene expression in the central nucleus of the amygdala is not specific for anxiolytic or anxiogenic drugs. Neuropharmacology, 50(1), 57-68. Show abstractHide abstract

The lateral, basal, and central nuclei of the amygdala are part of a circuitry that instantiates many fear and anxious behaviors. One line of support indicates that immediate-early gene (IEG) expression (e.g., c-fos and egr-1 (zif268)) is increased in these nuclei following fear conditioning. Other research finds that anxiogenic drugs working through various mechanisms induce IEG expression in the central nucleus of the amygdala (CeA) suggesting that expression is a neural marker for fear and anxiety. However, several studies have also found that anxiolytic drugs induce IEG expression in the CeA. Expression of egr-1 in the CeA and lateral nucleus of the amygdala following administration of anxiolytic and anxiogenic benzodiazepine and serotonin agonists and antagonists was investigated. The first experiment determined behaviorally active anxiolytic and anxiogenic doses for two anxiogenic drugs (FG 7142 and mCPP) and two anxiolytic drugs (diazepam and buspirone). The effects of anxiogenic and anxiolytic doses of these drugs on egr-1 expression in the amygdala were then tested in a second experiment. All four drugs increased egr-1 in the CeA indicating that increased egr-1 mRNA expression in the CeA is not specific to anxiolytic or anxiogenic effects of the drugs. We suggest that IEG expression in the CeA may be due to activation of circuits that are associated with systemic physiological homeostasis perturbed by a number of drugs including anxiogenic and anxiolytic compounds.

Thompson, B. L., Levitt, P., & Stanwood, G. D. (2005). Prenatal cocaine exposure specifically alters spontaneous alternation behavior. Behavioural Brain Research, 164(1), 107-116. Show abstractHide abstract

Our laboratory has previously characterized a rabbit model of gestational cocaine exposure in which permanent alterations in neuronal morphology, cell signaling and psychostimulant-induced behavior are observed. The cellular and molecular neuroadaptations produced by prenatal cocaine occur in brain regions involved in executive function and attention, such as the anterior cingulate and medial prefrontal cortices. Therefore, in the present study, we have measured the effects of prenatal cocaine exposure on specific behavioral tasks in adult offspring whose mothers were treated with cocaine (3mg/kg, twice a day, E16-E25). We assessed non-spatial, short-term memory in a two-object recognition task and found no deficits in memory or exploratory behaviors in cocaine-exposed offspring in this paradigm. We also evaluated a different memory task with a more robust attentional component, using spontaneous alternation in a Y maze. In this task, young adult rabbits exposed to cocaine prenatally exhibited a significant deficit in performance. Deficits in spontaneous alternation can be induced by a wide variety of behavioral and cognitive dysfunctions, but taken together with previous findings in this and other animal models, we hypothesize that prenatal exposure to cocaine alters highly specific aspects of cognitive and emotional development.

Rosen, J. B., Adamec, R. E., & Thompson, B. L. (2005). Expression of egr-1 (zif268) mRNA in select fear-related brain regions following exposure to a predator. Behavioural Brain Research, 162(2), 279-288. Show abstractHide abstract

Research has demonstrated that immediate-early genes/inducible transcriptional factors (e.g., c-fos, egr-1) are increased in amygdala nuclei (lateral, basal and central nuclei) known to be involved in fear conditioning, footshock stress and novelty. Although these data suggest that expression of inducible transcriptional factors are involved in fear, other non-shock ethologically based paradigms (predator or predator odor exposure) do not appear to increase c-fos in the lateral and basal nuclei. While the lack of c-fos expression may indicate that predator stress does not engage the lateral and basal amygdala nuclei, it may be that c-fos in the amygdala is not responsive to predator exposure. Therefore, egr-1, which increases in the lateral nucleus following fear conditioning, footshock and novelty, was assessed to determine if its expression is induced in rats exposed to a cat. Five minutes of cat exposure did not increase expression of egr-1 mRNA in the lateral nucleus of the amygdala. egr-1 was increased in the paraventricular nucleus of the hypothalamus, indicating cat-induced stress, and visual cortex compared to rats that were either confined for 5 min or handled. In the lateral periaqueductal gray, handled rats displayed a left hemisphere dominance, which disappeared in both the cat-exposed and confined group, suggesting that immobility, induced by either cat-induced stress or unstressed confinement, increased right hemisphere egr-1 expression. The results are discussed in a context of differences and similarities in neural circuitry for conditioned and unconditioned fear.

Thompson, B. L., Erickson, K., Schulkin, J., & Rosen, J. B. (2004). Corticosterone facilitates retention of contextually conditioned fear and increases CRH mRNA expression in the amygdala. Behavioural Brain Research, 149(2), 209-215. Show abstractHide abstract

The present study examined the effects of glucocorticoid administration on emotional memory and on corticotropin-releasing hormone (CRH) mRNA expression in the central nucleus of the amygdala (CeA) and the paraventricular nucleus of the hypothalamus (PVN). This was tested by administering repeated corticosterone (CORT) within a contextual fear conditioning paradigm. Rats received 2.5 mg/kg (s.c.) CORT or placebo twice a day for five and a half days and, 2 h after the last injection, rats were given one-trial contextual fear conditioning. When tested for retention of conditioned fear 6 days later, the CORT-treated rats displayed more fear-conditioned freezing in the retention test than vehicle-treated rats, which was not accounted for by an increase in footshock responsivity nor elevated plasma CORT. Another group of rats was fear conditioned prior to CORT administration, followed 24 h later by the five and a half days of CORT, and tested 6 days later; conditioned fear was not enhanced in these rats. Finally, CORT administration produced an increase of CRH mRNA in the CeA and a decrease in the PVN. The data suggest that repeated administration of CORT given before fear conditioning facilitates the acquisition of emotional memory, whereas CORT given after consolidation does not increase emotional memory.

Schulkin, J., Thompson, B. L., & Rosen, J. B. (2003). Demythologizing the emotions: Adaptation, cognition, and visceral representations of emotion in the nervous system. Affective Neuroscience, 52(1), 15-23. Show abstractHide abstract

This article highlights four issues about the neurobiology of emotions: adaptation vs. dysfunction, peripheral and central representations of emotion, the regulation of the internal milieu, and whether emotions are cognitive. It is argued that the emotions evolved to play diverse adaptive roles and are biologically vital sources of information processing. They were not designed as pieces of pathology, though they certainly can underlie some psychophathologies. Emotions are, in part, appraisal systems that are operative at numerous level of the nervous system from the brainstem to the cortex. Like other information processing systems they are not perfect cognitive systems. Emotional systems often utilize somatic and visceral information for appraisals of events to facilitate decisions of whether to approach or avoid objects. The neural systems of emotions traverse the entire neural axis and are linked to the regulation of the internal milieu. Thus, in addition to the experiential aspects of emotions, emotions embody appraisal systems that are pervasive to all levels of the brain to facilitate function, adaptation, and survival.

Markison, S., Thompson, B. L., Smith, J. C., & Spector, A. C. (2000). Time course and pattern of compensatory ingestive behavioral adjustments to lysine deficiency in rats. The Journal of Nutrition, 130(5), 1320-1328. Show abstractHide abstract

We and others have demonstrated that rats deficient in an essential amino acid (EAA) will consume sufficient quantities of the lacking nutrient to produce repletion when it is made available in solution. In the current series of experiments, we made rats deficient in lysine (LYS) by limiting the level of this EAA in the diet. We then examined licking behavior during approximately 23-h two-bottle intake tests over 4 consecutive days. In three separate experiments, rats were presented with the following: 1) 0.1 mol/L LYS and water, 2) 0.2 mol/L threonine (THR) and water and 3) 0.1 mol/L LYS and 0.2 mol/L THR. Lysine-deficient (LYS-DEF) rats drink significantly more LYS than did nondepleted controls (CON) when this amino acid was available. Meal pattern analysis revealed that the enhanced intake of LYS occurred as a function of a greater number of ingestive bouts, not changes in bout size. A cumulative analysis of LYS intake between CON and LYS-DEF rats revealed that a potentiation of intake developed within 30 min of sampling the solution when LYS and water were available and within 90 min when LYS and THR were the contrasting choices. In conclusion, increased LYS intake in the deficient rats occurs relatively rapidly and appears to be at least somewhat specific. Moreover, LYS deficiency does not seem to enhance the palatability of the limiting amino acid as judged by behaviors such as lick rate and bout size. Instead, LYS-DEF rats relieve the deficiency by increasing the number of drinking episodes initiated.

Thompson, B. L., & Rosen, J. B. (2000). Effects of TRH on acoustic startle, conditioned fear and active avoidance in rats. Neuropeptides, 34(1), 38-44. Show abstractHide abstract

The effects of intracerebroventricular injection of thyrotropine-releasing hormone (TRH) on acoustic startle, conditioned fear and active avoidance were examined in rats. Acoustic startle was significantly depressed by 12.5 microg TRH, while increasing motor activity. In a fear-potentiated startle paradigm, 12.5 microg TRH reduced the overall startle response amplitude, but did not decrease the amount of fear-potentiated startle. When TRH was administered 15 min before contextual fear conditioning, neither fear-related freezing in acquisition nor in a retention test was affected. In contrast, when TRH was administered 15 min before the retention test, TRH significantly reduced mean percentage of time spent freezing. TRH had no effect on active avoidance. The results demonstrate that TRH decreased acoustic startle and freezing responses, but had little effect on fear conditioning and active avoidance. It is suggested that the results may be due to TRH's effects on motor activity and arousal, independent of its effects on fear.