Objectives/Aims The bed nucleus of the stria terminalis (BNST), considered a part of ‘extended amygdala’ and located in the basomedial forebrain, is characterized by its connections with other limbic structures, hypothalamic and brainstem nuclei. With these crucial connections, it plays a key role in emotional processing and modulation of autonomic physiological and behavioural responses. Consequently, its dysfunction is implicated in many human psychiatric illnesses including depression thus positioning it as a target for neuromodulation with deep brain stimulation (DBS) and investigations of its direct neural activity.

Methods Here, in a cross-sectional design, we acquired task-induced neuronal recordings from BNST from fifteen treatment resistant depressed patients undergoing DBS surgery in a clinical trial. We employed two tasks: 1) empathy induction with painful and non-painful pictures, 2) affect task with emotionally valenced pictures (positive, neutral and negative). We first characterised task induced spectral dynamics and subsequently using permutation testing to assess condition differences. Time- frequency clusters which showed condition differences were then tested for their association with behavioural ratings and depression scores.

Results Empathy and affect task both showed task induced responses with condition differences predominantly in theta-alpha range. For empathy task, alpha activity also predicted the painful and non-painful behavioural ratings. Furthermore, alpha activity for non-painful pictures were also significantly associated with depression scores. Affect task showed condition differences in activity for positive and negatives pictures when compared to neutral pictures. Alpha activity of negative pictures (when compared to neutral pictures) were also positively correlated with depression scores.

Conclusions Overall, from direct BNST recordings, we thus show the spectral signatures of BNST in evaluation of empathy and affect pictures. Furthermore, alpha activity in these tasks also predicted depression scores thus highlighting its role as a potential biomarker for depression severity.