Background
Traumatic experiences early in life, such as child abuse, neglect and parental loss, are major risk factors for the development of a range of psychiatric and somatic disorders in adulthood, including depression, posttraumatic stress disorder (PTSD), and chronic fatigue syndrome (CFS) [
1‐
3]. There is ample evidence from animal models and human studies that adverse experience early in life, during periods of heightened brain plasticity, permanently programs the development of multiple brain circuits involved in the processing of environmental stimuli and the regulation of behavioral, autonomic, and endocrine responses to stress [
4]. Studies in humans have shown that several cardinal biological features of depression or CFS are secondary to early-life trauma and might reflect vulnerability for the development of these disorders [
5,
6].
Several of the disorders that have been linked to childhood trauma are characterized by profound cognitive impairment, including depression, PTSD, and CFS. There is direct evidence from animal models that early-life stress induces structural, functional, and epigenetic changes in brain regions involved in cognition, most notably the hippocampus [see [
4]]. Small hippocampal volume is associated with cognitive impairment, specifically memory deficits, in patients with depression or PTSD [
7‐
9]. While hippocampal impairment may be the result of toxic effects of chronic cortisol overexposure or increased glucocorticoid sensitivity [
4], more recent evidence suggests that small hippocampal volume might be a preexisting risk factor [
4,
4,
9,
10]. Of note, hippocampal volume loss in depression has been associated with early-life trauma [
11]. It might be conceivable that cognitive impairment in these disorders is linked to early-life trauma. Whether or not early-life stress is associated with long-term cognitive deficits in humans is largely unknown.
Maternal separation of rats results in enduring hippocampal dysfunction, including impaired memory and spatial learning [
12‐
16]. Juvenile rats exposed to a platform stress perform poorer on a spatial learning task in adulthood than control rats [
17]. These cognitive deficits persist through late adulthood and early aging [
18]. Similar results have been obtained in studies with non-human primates. Rhesus macaque monkeys reared in isolation exhibit significant deficits in learning and memory [
19‐
21]. Few studies have investigated associations between childhood adversity and cognitive function in humans. Infants and toddlers who have been abused, neglected or exposed to multiple medical and surgical procedures often exhibit deficits in cognitive, language, and motor skills [
22,
23]. Palmer et al. [
24] found profound impairments in cognition, including intellectual development delays and language and psychomotor deficiencies in sexually abused children. Abused or neglected children also have a higher risk for poor academic achievement [
25]. On the other hand, two studies comparing groups of trauma-exposed and non-exposed children and adolescents on cognitive outcome measures found that trauma exposure was not associated with lower estimates of intelligence [
26] or memory and learning deficits [
27]. Given the limited information and inconsistency of the human literature, more human research is needed that examines cognitive performance of persons exposed to traumatic events early in life.
The aim of this study was to assess the association between childhood trauma exposure and cognitive function in healthy adults. We hypothesized that exposure to childhood trauma would be significantly associated with impairment in cognition, specifically hippocampus-related memory function, and that childhood trauma would be significantly associated with academic underachievement. To test these hypotheses we measured childhood trauma exposure, neurocognitive function, and level of academic achievement in a group of healthy adults (without concurrent psychiatric illness) randomly selected from the Wichita, KS population. These subjects were recruited as controls for a larger CDC study [
28]. We intended this pilot study to provide impetus for the development of future case-control matched studies that further scrutinize the association between childhood stress and adult cognitive function.
Results
Sample characteristics are shown in Table
1. As can be seen in the Table, mean CTQ scores across all subjects were in the none-mild range. Nevertheless, a proportion of 25.5% of subjects had experienced at least one type of maltreatment that scored above the cut-off for moderate-severe abuse. The most common form of reported maltreatment was emotional abuse. Mean depression ratings were in the normal range (i.e. no depression) and state and trait anxiety ratings were low (see Table
1). None of the subjects had a depression score in the clinically significant range; none of the subjects had state or trait anxiety above the population average.
Table 1
Sample Characteristics and Early Life Stress (n = 47)
Age (SE; range) | 51.51 (1.22; 31-69) |
Sex | 7 male, 40 female |
Race | 44 white, 2 black, 1 other |
Education | |
≤ High school or vocational tech diploma | 20 |
Associate degree, RN diploma or college | |
Income ($/year)a
| 27 |
20,000 - 40,000 | |
> 40,000 | 13 |
WRAT-3 standard score (SE) | 33 |
CTQ Total score (5-125) (Mean (SE)) | 33.7 (1.34) |
Emotional abuse (5-25) (Mean (SE)) | 7.7 (.516) |
Physical abuse (5-25) (Mean (SE)) | 6.36 (.259) |
Sexual abuse (5-25) (Mean (SE)) | 5.85 (.373) |
Emotional neglect (5-25) (Mean (SE)) | 7.9 (.479) |
Physical neglect (5-25) (Mean (SE)) | 5.85 (.228) |
CTQ Exposures b, n (%) | 12 (25.5%) |
Emotional abuse | 8 (17.02%) |
Physical abuse | 4 (8.51%) |
Sexual abuse | 3 (6.38%) |
Emotional neglect | 2 (4.26%) |
Physical neglect | 2 (4.35%) |
Self-Rating Depression Scale (Mean (SE)) | 37.5 (.106) |
Spielberger State Anxiety (Mean (SE)) | 26.1 (.832) |
Spielberger Trait Anxiety (Mean (SE)) | 27.5 (.772) |
Achievement
After adjustment for age, sex, education, and income, multiple linear regression analysis revealed an association between the sexual abuse and the physical neglect scores of the CTQ and the WRAT-3 standard score (sexual abuse: adjusted B = -6.18, SE = 2.86, p = 0.03; physical neglect: adjusted B = -9.28, SE = 3.97, p = 0.02). However, the association did not achieve a more conservative statistical significance level of 0.01 for multiple separate hypothesis testing. The association indicated that more exposure to sexual abuse or physical neglect in childhood was associated with worse performance in the WRAT-3 test as an adult.
Memory
Multiple linear regression analyses revealed three significant associations between CTQ scale scores and CANTAB measures of memory (Table
2). The emotional abuse score was significantly associated with the number of double errors in the Spatial Working Memory test. Furthermore, there was a significant association between the physical neglect score and the number of double errors in the Spatial Working Memory test and the latency for a correct response in the Pattern Recognition Memory test. The more exposure to these two different types of childhood trauma, the worse was the memory performance.
Table 2
Associations of exposure to childhood trauma with CANTAB measures of memory
Spatial Working Memory
| | | | | | |
Strategy score | 34.18 (0.53) | -0.07 (0.86) | -0.23 (1.80) | 0.74 (1.12) | -0.25 (0.93) | 0.06 (1.84) |
Standardized B
c
| | -0.01 | -0.03 | 0.11 | -0.05 | 0.00 |
# Double errors (total) | 2.74 (0.74) | 3.44 (1.05)** | 5.92 (2.30)* | 0.56 (1.54) | 0.80 (1.27) | 6.59 (2.31)** |
Standardized B
c
| | 0.48 | 0.41 | 0.07 | 0.10 | 0.40 |
Spatial Recognition Memory
| | | | | | |
% Correct | 84.15 (1.08) | -0.31 (1.72) | -4.38 (3.55) | 0.91 (2.25) | -0.25 (1.87) | 0.01 (3.69) |
Standardized B
c
| | -0.03 | -0.21 | 0.06 | -0.02 | 0.00 |
Mean correct latency (ms) | 2567.11 (110.29) | -39.01 (170.51) | -225.73 (357.48) | -129.08 (222.88) | 32.12 (185.40) | 393.77 (361.68) |
Standardized B
c
| | 0.31 | -0.11 | -0.09 | 0.03 | 0.16 |
Pattern Recognition Memory
| | | | | | |
% Correct | 90.34 (1.48) | -5.49 (2.20)* | -9.90 (4.73) | 4.33 (3.03) | -1.29 (2.56) | -2.54 (5.08) |
Standardized B
c
| | -0.39 | -0.35 | 0.22 | -0.08 | -0.08 |
Mean correct latency (ms) | 2304.76 (123.25) | 460.12 (180.88) | 717.64 (394.50) | -187.78 (254.00) | 238.08 (208.55) | 1213.18 (372.90)** |
Standardized B
c
| | 0.39 | 0.30 | -0.11 | 0.19 | 0.45 |
Executive Functions
As shown in Table
3, there were no significant associations found between CTQ scale scores and CANTAB measures of executive functions. The two measures of the Intra/Extra Dimensional Shift task, assessing rule acquisition and reversal, featuring visual discrimination and shifting of attention, were not statistically significantly associated with any of the CTQ scale scores. Similarly, there were no significant associations between the different CTQ scales and reasoning and planning abilities, as measured by the Stockings of Cambridge task.
Table 3
Associations of exposure to childhood trauma with CANTAB measures of executive functions
Intra/Extra Dimensional Shift
| | | | | | |
# Stages completed | 8.43 (0.14) | 0.05 (0.23) | -0.53 (0.48) | -0.73 (0.28) | -0.35 (0.25) | -0.71 (0.49) |
Standardized B
c
| | 0.04 | -0.19 | -0.38 | -0.24 | -0.23 |
# EDS errors | 9.66 (1.44) | -0.68 (2.36) | 1.90 (4.96) | 0.01 (3.10) | -1.32 (2.56) | 0.81 (5.08) |
Standardized B
c
| | -0.05 | 0.07 | 0.00 | -0.09 | 0.03 |
Stockings of Cambridge
| 8.83 (0.31) | -0.42 (0.45) | -1.16 (0.95) | 0.93 (0.58) | -0.16 (0.50) | 0.52 (0.98) |
Total # perfect solutions | | -0.14 | -0.19 | 0.23 | -0.05 | 0.08 |
Standardized B
c
| 5.37 (0.14) | 0.23 (0.19) | 0.30 (0.41) | -0.11 (0.26) | 0.23 (0.21) | 0.08 (0.42) |
4-Move Problems | | 0.17 | 0.11 | -0.06 | 0.16 | 0.03 |
Average # moves | 11624.73 | -2311.04 | -3940.05 | -748.42 | -2716.01 | 1787.55 |
Standardized B
c
| (1055.29) | (1423.33) | (3030.73) | (1923.94) | (1538.53) | (3148.11) |
ITT (ms) | | -0.24 | -0.20 | -0.06 | -0.26 | 0.08 |
Standardized B
c
| 2071.30 (321.45) | 106.29 (460.81) | -551.89 (966.98) | 36.82 (604.83) | -6.59 (501.23) | 212.98 (991.26) |
STT (ms) | | 0.03 | -0.09 | 0.01 | -0.00 | 0.03 |
Standardized B
c
| | | | | | |
5-Move Problems | 6.40 (0.20) | 0.28 (0.31) | 0.92 (0.65) | -0.67 (0.40) | -0.01 (0.34) | -0.17 (0.68) |
Average # moves | | 0.14 | 0.23 | -0.25 | -0.00 | -0.04 |
Standardized B
c
| 13879.70 | -1568.09 | -8073.42 | -230.52 | 621.74 | -115.66 |
ITT (ms) | (1389.96) | (2185.76) | (4452.37) | (2885.29) | (2389.13) | (4731.53) |
Standardized B
c
| | -0.12 | -0.30 | -0.01 | 0.04 | -0.00 |
STT (ms) | 1021.84 (161.05) | 210.91 (253.29) | 626.28 (528.72) | -498.01 (325.73) | -199.99 (275.89) | -212.60 (548.48) |
Standardized B
c
| | 0.14 | 0.20 | -0.23 | -0.12 | -0.06 |
Psychomotor Speed and Sustained Attention
Performance on the Reaction Time task, as measured by five-choice reaction and five-choice movement times, was not significantly associated with the CTQ scales, nor were there significant associations between the CTQ scale scores and performance as measured by signal A' and mean latencies to correct responses in the Rapid Visual Information Processing test (Table
4).
Table 4
Associations of exposure to childhood trauma with CANTAB measures of psychomotor speed and sustained attention
Reaction Time
| | | | | | |
5-choice reaction time (ms) | 396.20 (8.66) | -3.86 (14.07) | -16.38 (29.53) | -7.07 (18.44) | -15.34 (15.11) | 14.16 (30.20) |
Standardized B
c
| | -0.05 | -0.10 | -0.06 | -0.17 | 0.07 |
5-choice movement time (ms) | 479.97 (16.28) | 60.21 (24.31) | 70.17 (53.84) | -34.80 (33.80) | 19.63 (28.21) | 95.84 (54.07) |
Standardized B
c
| | 0.39 | 0.23 | -0.16 | 0.12 | 0.27 |
Rapid Visual Information Processing
| | | | | | |
Mean latency (ms) | 548.43 (19.58) | 7.06 (27.89) | 33.45 (58.87) | -31.36 (36.28) | 15.57 (31.28) | 40.50 (59.82) |
Standardized B
c
| | 0.04 | 0.09 | -0.12 | 0.08 | 0.10 |
A' | | -0.016 (0.010) | -0.023 (0.022) | -0.016 (0.013) | -0.026 (0.011) | -0.021 (0.022) |
Standardized B
c
| 0.90 (0.01) | -0.25 | -0.17 | -0.18 | -0.37 | -0.14 |
Discussion
This pilot study found significant associations between level of childhood trauma exposure and cognitive performance in CANTAB measures of long-term and working memory in a group of healthy adults, with no significant symptoms of depression or anxiety, that were randomly selected from the general population. Healthy adults with high exposure to emotional abuse, the most common form of reported maltreatment in this sample, exhibited a higher error rate in the Spatial Working Memory test. Furthermore, individuals with high levels of exposure to physical neglect showed a higher error rate in the Spatial Working Memory test and prolonged latency to a correct response in the Pattern Recognition Memory test. Finally, we found a (less significant) association between level of exposure to sexual abuse or physical neglect and lower scores in the reading subtest of the WRAT-3, indicating less academic achievement in traumatized subjects. Our results add to a growing literature that supports a relationship between childhood trauma exposure and the development of cognitive dysfunction in children and poor academic achievement [
22‐
25]. The current findings suggest that memory deficits are specifically associated with childhood trauma exposure in healthy adults.
Negative associations between childhood trauma exposure and cognitive performance were found in the domains of long-term and working memory. Working memory refers to the structures and processes used for temporarily storing and manipulating information. Long-term memory differs structurally and functionally from working memory. It holds information from a few minutes to decades [
37]. Participants with higher levels of physical neglect showed longer response latencies in the Pattern Recognition Memory task, a test for long-term memory. This deficit cannot be explained by reduced elemental speed of cognitive processing, since there was no significant association between physical neglect and the response time measures in the Reaction Time task. Therefore, our results suggest a specific deficit in the ability to judge the prior occurrence of visual patterns (long-term memory) in subjects with higher levels of exposure to physical neglect. In Spatial Working Memory, we found that subjects with more exposure to physical neglect or emotional abuse had a higher rate in double errors. Efficient solving of problems in the spatial working memory test requires remaining highly attentive, using memory skills to remember previously selected and targeted locations, and developing and maintaining strategies to organize each search. Attentional problems did not appear to affect performance in Spatial Working Memory since there was no association between level of physical neglect or emotional abuse and performance in the Rapid Visual Information Processing task, a test of sustained attention. Organizational abilities were also intact in subjects with higher levels of childhood trauma since we found no relationship between the strategy scores in the Spatial Working Memory task or the scores in the Stockings of Cambridge task and physical neglect or emotional abuse. Therefore the higher error rate in the Spatial Working Memory task in subjects with higher levels of physical neglect or emotional abuse presumably reflects a pure memory deficit in these subjects. Although no brain activity was measured in this study, one might speculate that the observed memory deficits are linked to altered structure of function of brain regions. Animal models of early-life stress suggest that medial temporal structures are affected. Rats, exposed to a period of early-life stress, show late-onset, selective deterioration of cognitive performance in a hippocampus-involving object recognition memory task [
8]. Clinical evidence is brought by studies of patients with hippocampal damage. Patients with temporal lobe or amygdala-hippocampectomy damage show cognitive deficits in pattern recognition memory compared to those with frontal lobe excision [
35]. Furthermore, patients suffering from senile dementia of the Alzheimer type show impairments in spatial working memory that are accompanied by evidence of an intact strategy approach to the task, suggesting a pure memory deficit associated with hippocampal impairment in these patients [
38]. Of note, memory deficits are core features of depression and PTSD, and these disorders are also associated with hippocampal damage [
4,
7‐
9]. Hippocampal volume loss in depression has been associated with early life trauma [
11]. Taken together, our study results suggest that exposure to emotional neglect or physical abuse is associated with cognitive underperformance in tasks that involve the hippocampus.
Cognitive deficits after childhood trauma could be a direct consequence of the effects of trauma on the brain or could occur as a result of psychiatric illness, alcohol and substance abuse, or medical illness, which are associated with childhood trauma. Especially, the deficits in working and long-term memory found in this study are quite characteristic for patients with depression [
4,
7]. In our study, however, cognitive deficits were found in adult survivors of childhood trauma who did
not suffer a current medical or psychiatric illness (including major depressive disorder) and did
not have a history of alcohol or substance abuse. Subjects were also free of subsyndromal depression and anxiety as tested using rating scales. Hence, the cognitive deficits linked to childhood trauma are not secondary to depression or other psychiatric or medical illnesses.
There are several limitations of the present study. First, sample size is very small, which might have led to false positive or false negative results due to outliers. Replication of our findings in larger samples is necessary. Such studies might include subjects recruited based on childhood trauma in order to obtain more cases with severe trauma and larger cells for different trauma types, which would allow for comparisons of cognitive function between groups. Second, we relied on retrospective and uncorroborated self-reports of childhood experiences. Problems concerning the credibility of self-reported childhood trauma include simple forgetting, non-awareness, nondisclosure, and reporting biases due to mood states. However, the use of validated psychometric instruments increases validity of self-reported data [
39]. Third, we did not consider effects of adulthood trauma and life stress that might mediate or moderate the relationship between childhood adversity and cognitive dysfunction, since individuals with early adverse experience more frequently experience adulthood stresses, and moreover, are sensitized to the effects of such stressors [
40]. Fourth, owing to the cross-sectional design, we cannot determine whether cognitive dysfunction might have preceded early adversity. Finally, it must be noted that we studied a sample comprising subjects who had significant exposure to early-life trauma but remained healthy. Indeed, there is a substantial amount of resilience after early-life stress [
41,
42]. One must consider the possibility that our findings reflect cognitive changes specific to resilient persons with early life trauma, which might be different from those persons with early-life trauma who go on to develop a disorder.
Conclusions
Results of this pilot study should be considered as preliminary. Our observations lend support for the hypothesis that exposure to childhood trauma, especially emotional abuse and physical neglect, leads to problems in long-term and working memory in adulthood. Longitudinal studies are needed to provide more information on the causal relationship between childhood trauma and memory deficits, and to systematically evaluate developmental trajectories as well as mediators and moderators of this relationship. These studies might consider maternal education and mother-child interactions as moderator variables as developmental psychologists have reported that these variables are partially predictive of the cognitive performance of children [
43]. Future studies might also consider the potential role of socioeconomic status in contributing to the cognitive deficits. Our results might have important clinical implications as cognitive impairment is a key factor in many psychiatric disorders, such as major depression or post-traumatic stress disorder. Cognitive deficits should be addressed in the treatment of victims of childhood trauma, independent of their being concurrently diagnosed with a psychiatric condition. Neuropsychological training should be implemented in a comprehensive intervention strategy for individuals with childhood trauma. Furthermore, early intervention may prevent the long-term deficits in memory function, and hence academic performance in maltreated children.
Acknowledgements
We thank Abt Associates, Inc., for their contributions. The study was fully funded by the US Centers for Disease Control and Prevention and conducted in collaboration with Emory University School of Medicine. The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the funding agency. The authors have no financial interests related to the results of this study.
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the funding agency.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors read and approved the final manuscript. WCR was principal investigator of the study. MM, UMN, JSL, LC, and WCR wrote the manuscript. All authors contributed to the manuscript. WCR designed the study protocol and supervised data collection during the study. WCR supervised the conduct of clinical studies and data collection. MM and JSL conducted the statistical analyses.