Sensitive periods in development and risk for psychiatric disorders and related endpoints: a systematic review of child maltreatment findings

Abstract

Variation in the mental health of people who have experienced childhood maltreatment is substantial. One hypothesis is that this variation is attributable, in part, to the timing of maltreatment-specifically, whether maltreatment occurs during sensitive periods in development when the brain is maximally sensitive to particular types of environmental input. To determine whether there is scientific consensus around when periods of peak sensitivity occur, we did a systematic review of human observational studies. Although 89 (75%) of the 118 unique cross-sectional or longitudinal cohort studies we identified reported timing effects, no consistent sensitive periods were identified for any of the most studied outcomes. Thus, observational research on childhood maltreatment has yet to converge on a single period (or set of periods) of increased vulnerability. We identified study characteristics that might contribute to these between-study differences and used observations from our Review to suggest a comprehensive set of recommendations for future research.

Figure 1.. Study selection flow diagram for our systematic review

Figure 2.. Overview of the existing literature on sensitive periods for child maltreatment

Note. This figure summarizes different characteristics of the 118 studies presented in Table S2, color-coded by study level of analysis. (A) Scatterplot showing increases in sample size across sensitive period studies over time. (B) Frequency of child maltreatment exposures examined across sensitive period studies. Studies could be counted more than once in this panel if they examined multiple maltreatment types. “Maltreatment composite” refers to studies that combined multiple types of maltreatment into a single exposure measure. “Adversity composite” refers to studies that combined maltreatment and other non-maltreatment adversities into a single exposure measure. (C) Frequency of specific outcomes examined across sensitive period studies. Studies could be counted more than once in this panel if they examined multiple outcomes. (D) Operationalization of exposure timing across sensitive period studies. Researchers inferred the occurrence of sensitive periods based on findings related to the effects of maltreatment reported at different time points in development, with timing based on: (1) age at first or worst exposure; (2) presence/absence of maltreatment during different windows of development (e.g., between ages 0–11 vs. 12–18, or each year between ages 0–18); (3) scores capturing the extent of exposure during different windows of development (e.g., frequency, severity, total number of maltreatment events, or number of types of maltreatment), or (4) grouping by recency of maltreatment, often combined with maltreatment onset (e.g., comparing an “early-maltreated” to “early-and-recent” and/or “recent-only” maltreated groups). This figure displays the distribution of ways in which these various definitions of “maltreatment age” were operationalized across sensitive period analyses, including dividing age into 2–6 categories or by examining year of age (“continuous”). (E) Frequency of analytic strategies used across sensitive period studies. Indirect comparison studies (n=41; 34.7%) refer to those that describe, rather than directly test, differences in the magnitude of the association between exposure and outcome as a function of exposure timing. Direct comparison studies (n=55; 46.6%) provide a stronger test of the sensitive period hypothesis, because they empirically evaluate whether the effect of exposure at one point in development differs from the effect of exposure at another. Competing hypotheses studies (n=20; 16.9%) are the strongest test, because they directly evaluate alternative theoretical explanations, including accumulation, chronicity, and/or recency. Two studies (1.7%) did not fall cleanly into any of these three categories and thus are not shown.

Figure 3.. A summary of maltreatment-related sensitive period findings by level of analysis.

Notes. This figure summarizes results from studies testing for timing-of-child-maltreatment effects on the most commonly-studied outcome for each of three different levels of analysis: (1) depressive symptoms or diagnoses, (2) hippocampal volume, and (3) tests of cognitive or language ability. Studies are organized by their operationalization of maltreatment timing (e.g., dichotomous, categorical, or continuous). Each line depicts the results from one study, with the length and position of the line showing the ages when maltreatment was assessed. Thicker, color-coded line segments indicate the specific sensitive periods identified by the study, if any. Solid black dots denote the end of the maltreatment exposure-assessment window. Open black dots indicate studies where maltreatment exposure was assessed beyond age 18. Colored gradients depict sensitive periods from studies using a continuous operationalization of age and reported that outcomes were associated with an “earlier” or “later” sensitive period (i.e., without denoting specific ages) (see (57) as an example).

Figure 4.. Maltreatment-related sensitive period studies for depression, arranged by (a) sample size, (b) analytic strategy, (c) maltreatment assessment approach, (d) study design, and (e) diagnosis vs. symptom outcomes.

Notes. This figure summarizes results from studies testing for timing-of-child-maltreatment effects on depressive symptoms or Major Depressive Disorder diagnosis. Each line depicts the results from one study, with the length and position of the line showing the ages when maltreatment was assessed. Thicker line segments indicate the specific sensitive periods identified by the study, if any. Solid black dots denote the end of the maltreatment exposure-assessment window. Open black dots indicate studies where maltreatment exposure was assessed beyond age 18. Colored gradients depict sensitive periods from studies using a continuous operationalization of age and reported that outcomes were associated with an “earlier” or “later” sensitive period (i.e., without denoting specific ages) (see (57) as an example).

Figure 4.. Maltreatment-related sensitive period studies for depression, arranged by (a) sample size, (b) analytic strategy, (c) maltreatment assessment approach, (d) study design, and (e) diagnosis vs. symptom outcomes.

Notes. This figure summarizes results from studies testing for timing-of-child-maltreatment effects on depressive symptoms or Major Depressive Disorder diagnosis. Each line depicts the results from one study, with the length and position of the line showing the ages when maltreatment was assessed. Thicker line segments indicate the specific sensitive periods identified by the study, if any. Solid black dots denote the end of the maltreatment exposure-assessment window. Open black dots indicate studies where maltreatment exposure was assessed beyond age 18. Colored gradients depict sensitive periods from studies using a continuous operationalization of age and reported that outcomes were associated with an “earlier” or “later” sensitive period (i.e., without denoting specific ages) (see (57) as an example).

Figure 5.. Sensitive periods identified by studies assessing 2+ different types of maltreatment and at least one positive sensitive period finding.

Note. This figure displays results from studies that tested for sensitive period effects influencing the relationship between multiple individual forms of maltreatment and psychopathology outcomes in the same sample, organized by operationalization of timing (e.g., dichotomous, categorical, or continuous). Results of each study are depicted as a group of colored lines, with the length and position of the lines showing the ages at which maltreatment was assessed and color denoting maltreatment type. Thicker segments indicate the specific sensitive periods identified for each outcome, if any, whereas thinner segments indicate no sensitive periods were identified for that age range. Solid black dots denote the end of the exposure-assessment window. Gradients depict sensitive periods from studies using a continuous operationalization of age and reported that worse outcomes were associated with an “earlier” or “later” sensitive period (i.e., without denoting specific ages).