Does childhood interpersonal trauma alter cortical excitation-inhibition balance in adults?

The study asks how trauma relates to cortical E/I balance

The paper asks a specific question: in young adults without current psychiatric disorders, is a history of childhood interpersonal trauma accompanied by a measurable change in cortical excitation (E) and inhibition (I) balance? Earlier work on trauma and brain function often included people with diagnosed mental disorders, making it hard to tell whether neural differences came from trauma itself or from comorbidity. This study excluded people with prior diagnoses, a history of psychopharmacological treatment, or neurological disease (including epilepsy, traumatic brain injury, and neuroinfection), recruiting only from the general population.

The sample comprised 80 people aged 19–25, grouped by Childhood Trauma Questionnaire (CTQ) total score: a low-trauma group of 41 with CTQ total ≤36, and a high-trauma group of 39 with CTQ total ≥51, corresponding to moderate-to-severe exposure. The groups did not differ significantly in age, sex proportion, or years of education. All participants also completed the State-Trait Anxiety Inventory (STAI) and the Beck Depression Inventory (BDI-II), used to exclude clinically relevant depression and anxiety.

The 1/f slope and the E/I ratio are related but distinct measures

The study describes E/I balance with two measures. The first is the aperiodic 1/f slope of the EEG power spectrum: under the paper's direction convention, a smaller absolute slope (a flatter spectrum) corresponds to increased excitation, while a steeper slope corresponds to increased inhibition. The second measure comes from modeling: the researchers first simulated a local cortical circuit of leaky integrate-and-fire (LIF) neurons, in which excitatory and inhibitory populations connect through four synaptic weights (JEE, JEI, JIE, JII), running two million simulations that systematically varied these weights.

The simulated firing activity was converted into population-level brain-like signals through a kernel method derived from a multicompartment network model, and 1/f slopes were extracted from those signals. A ridge regression model was then trained on the simulated data as an inverse mapping, converting measured EEG slopes into estimates of the E/I ratio. The E/I ratio is therefore not a directly measured physiological quantity but a value inferred from slopes through a model; agreement between the two measures shows the conclusion is not sensitive to the method choice, but they are not two independent pieces of biological evidence.

Resting posterior cortex shifts toward excitation

In the 5-minute eyes-closed resting recording, the high-trauma group had significantly lower absolute 1/f slope values than the low-trauma group, with effect size |d|>0.2 and p<0.05, concentrated over posterior electrodes. Applying the inverse model to the same data, the high-trauma group showed a significant increase in the inferred E/I ratio in the same posterior regions. Both analysis paths point the same way: the activity balance in posterior cortex moved toward excitation.

As a control, when frontal electrodes were pooled, neither the resting 1/f slope nor the inferred E/I balance showed significant group differences. This bounds the spatial scope of the resting-state conclusion: the change appeared posteriorly, not uniformly across the brain.

Task differences appear in stimulus-driven dynamics

The visual reaction-time task was split into pre-stimulus and post-stimulus segments, with between-group comparisons and within-group pre-to-post comparisons run separately. Direct between-group comparisons yielded few results: only a small number of electrodes in the long-interval block showed significant differences in slope or inferred E/I ratio, and most regions showed no effect.

Within-group pre-to-post comparisons revealed a clear pattern. The low-trauma group showed a widespread steepening of the slope after stimulus onset, with the inferred E/I ratio shifting in parallel toward inhibition; the high-trauma group lacked this frontal modulation. The paper interprets this as attenuated stimulus-related inhibition in the high-trauma group. Behaviorally, the high-trauma group had significantly lower accuracy in the short-interval block (0.97 vs 0.98) and significantly higher values on the reaction-time variability index sigma in both blocks, but mean reaction time and the tau parameter did not differ significantly, indicating mild, detectable differences rather than obvious task failure.

Current affective symptoms do not explain the E/I differences

The high-trauma group scored significantly higher on state anxiety, trait anxiety, and depression, which is itself a common finding. But correlations between inferred E/I ratio and these concurrent symptoms were weak: at rest, E/I ratio correlated with state anxiety at Spearman r=0.05 (q=0.056), with trait anxiety at r=0.13 (q<0.001), and with BDI-II at r=0.12 (q<0.001). The authors therefore argue that the between-group E/I differences cannot be simply attributed to current affective burden.

Correlation strength and statistical significance need to be kept apart. The association between CTQ total score and resting inferred E/I ratio was the only relationship showing a clear monotonic trend, yet r was only 0.25 (FDR-adjusted q<0.001). Trauma exposure and E/I shift move in the same direction, but the explanatory power at the individual level is limited; statistical significance does not license judgments about individuals.

The cross-sectional design cannot establish causation or stability

The study is cross-sectional, so it cannot determine whether the E/I changes are enduring developmental consequences of childhood trauma or adaptive responses that evolve over time. Trauma exposure was obtained through retrospective self-report, which may carry recall bias, and threat-type versus neglect-type adversity could not be separated because the CTQ yields a composite total score.

There are also measurement boundaries. Although the 1/f slope is widely used as an E/I marker, the paper notes that some studies failed to confirm the expected association with direct measures of GABAergic and glutamatergic function, and slope alone may conflate distinct neural mechanisms. The modeling framework is a local cortical circuit model that does not explicitly incorporate long-range cortico-cortical connectivity. These limits bound the strength of the conclusion: it supports "trauma history is associated with an E/I shift," not "trauma causes E/I change" or "this marker can diagnose individuals."

The key distinction is the division of labor across conditions

The most valuable distinction in this paper is not the single sentence "the trauma group has higher E/I," but how results divide across the two conditions: at rest, the high-trauma group showed a stable excitatory shift in posterior cortex that correlated only weakly with concurrent anxiety and depression; during the task, direct pre-stimulus differences between groups were few, and the real difference appeared in stimulus-driven inhibitory modulation, which was clear in the low-trauma frontal cortex and absent in the high-trauma group.

So comparing only task-state group levels would likely yield a "no difference" conclusion; only by including pre-to-post-stimulus change does the reduced inhibitory flexibility of the high-trauma group become visible. The authors describe this combination as a possible transdiagnostic neurophysiological vulnerability feature rather than a disease-specific biomarker, and explicitly note this remains speculative, since the study did not assess perceptual disturbances and had no longitudinal follow-up. For readers, the conditions and regions should take priority when judging this evidence: an excitatory shift in resting posterior cortex plus absent frontal inhibitory modulation during the task is the picture the paper actually reports.

Sources used in this answer