Fig. 4: Reduced decodability of ΔES from MEG activity linked to OC symptoms.

a, ΔES can be decoded from MEG activity using an iterative multivariate lasso regression (left y axis: decodability, measured as the correlation between trained and tested data at each timepoint in the epoch, plotted in blue as group-level mean decodability, with shaded area representing the standard error of the mean, N = 105; horizontal lines denote significant time clusters (P < 0.05 in a two-sided cluster-based permutation test at threhold t > 3.1)). Individual differences in OC factor predict significantly decreased decodability within 420–470 ms and 530–560 ms post stimulus (right y axis: beta coefficient for the OC factor predicting the decodability of ΔES at each timepoint in the epoch from the 3 factor scores in a GLM per timepoint, plotted in orange, N = 105; horizontal lines denote significant time clusters (P < 0.05 in a two-sided cluster-based permutation test at threshold t > 2.1)). b, Illustration of the attenuated ΔES decodability within 420–470 ms (N = 105), plotted here by categorical group, that is, for participants from the general population with low and high obsessive–compulsive scores (‘low C’ and ‘high C’, respectively), healthy controls, participants with GAD and participants with OCD. Note that this is merely a visualization of the significant difference found in a, showing individual data points. c, To investigate the spatial distribution of the sensors contributing to the decodability of ΔES within the time windows where we find individual differences (420–470 ms post stimulus onset), we used searchlight analyses (N = 105). ΔES are represented in a wider network comprising frontal, mediocentral and occipital areas (left). The attenuation of ΔES representations in high-OC participants was primarily driven by mediofrontal sensors, supporting the idea of altered ΔES processing in mediofrontal areas in high-OC participants (right).

Source data