Gamma oscillations in basolateral amygdala as a mechanistic and predictive biomarker for prefrontal DBS analgesia

Deep brain stimulation (DBS) targeting the prefrontal cortex (PFC) is a candidate strategy for pain modulation. However, its inconsistent efficacy highlights the need for closed-loop adaptive systems, which require neural biomarkers. The basolateral amygdala (BLA), a downstream target of the PFC, is involved in pain, yet the network oscillatory signatures that reflect and predict PFC-DBS analgesia remain elusive.

To determine how PFC-DBS modulates downstream BLA oscillatory activity during nociceptive processing, and to identify oscillatory features that predict stimulation efficacy.

In an acute pain model in rats, we applied PFC-DBS with varying stimulation parameters while simultaneously recording local field potentials in the basolateral amygdala (BLA). We analyzed the oscillatory response to nociceptive stimuli, examined DBS-induced changes in ongoing activity and applied a machine learning method to evaluate predictive biomarker features.

PFC-DBS showed a parameter-dependent analgesic effect accompanied by reduced nociceptive gamma power in the BLA. Nociceptive gamma activity was identified by a machine learning classifier as a predictive neural feature associated with analgesic outcomes. Crucially, DBS-induced ongoing oscillatory activity significantly distinguished effective from ineffective DBS-analgesia, with higher fast gamma power as top contributing feature for successful analgesia. The analgesic effect of enhanced ongoing fast gamma is partially mediated by reduced nociceptive gamma. Furthermore, the baseline oscillatory state before DBS predicted analgesia outcome.

We identify a state-response BLA gamma signature associated with effective PFC-DBS, characterized by enhanced ongoing fast gamma and suppressed nociceptive gamma activity. These findings provide neurophysiological support for developing closed-loop, adaptive DBS systems.