Aperiodic Subthalamic Dynamics Distinguish Rest from Affective Processing: Implications for Adaptive DBS
S. Karim, V. Voon, S. SonkusareIntroduction
The subthalamic nucleus (STN) is central to basal ganglia circuits, integrating motor, cognitive, and limbic functions. Its aberrant activity underlies core motor symptoms of Parkinson’s disease (PD). While its motor role is well established, evidence implicates the STN in affective processing. Deep brain stimulation (DBS) alleviates PD motor symptoms and, perioperatively, allows intracranial recordings from implanted electrodes. These recordings provide a unique opportunity to study STN dynamics across brain states. Neural signals contain oscillatory components and a non-oscillatory aperiodic component in the background. This aperiodic component follows a 1/f distribution, with two features: the exponent , the slope reflecting the level of local excitation–inhibition balance , and the offset , the overall broadband power (see Img. 1). Despite detailed STN oscillation studies, the aperiodic component remains under-investigated during emotion processing.
Objectives
We aimed to characterise STN aperiodic features—the exponent and offset—during rest and emotional stimulus engagement, and assess their potential as markers of state-dependent excitation–inhibition balance.
Methods
We analysed bilateral STN recordings from 19 PD patients (6 female; mean ± SD age = 59.7 ± 11.8 years) undergoing DBS. Recordings were obtained during rest and presentation of emotionally salient images from the International Affective Picture System (IAPS). Data were visually inspected and cleaned, then processed using the FOOOF algorithm to isolate the aperiodic exponent and offset. This yielded 114 electrode-level recordings across patients, which were analysed at the group level using paired t-tests (α = 0.05) between resting and task conditions.
Results
The aperiodic exponent was lower during emotional processing
(1.44 ± 0.04)
than at rest
(1.49 ± 0.04
; t(113) = 1.68,
p = 0.03
). The offset was also lower during the task
(0.159 ± 0.059)
than at rest
(0.211 ± 0.053)
, but this was not significant (t(113) = 2.18,
p = 0.09
).