Neural Information in Disease States
We apply information theory to directly quantify how neural circuits encode and transmit data in both healthy and disease states. By analyzing single-unit and network-level recordings, we have demonstrated that parkinsonism is characterized by pathological increases in neuronal entropy and aberrant information transmission. Crucially, our work reveals that Deep Brain Stimulation (DBS) effectively reduces this chaotic signaling, regularizing neural activity and restoring normalized information flow throughout the basal ganglia and thalamic networks.
Physiological Signal Analysis Methods
Accurately quantifying biological signal statistics requires robust computational methodologies, particularly when working with the limited datasets typical of in vivo electrophysiological recordings. To overcome these analytical challenges, we develop specialized algorithms that yield highly robust estimates from sparse data, e.g., of neural entropy and information from neuronal spike trains. Beyond novel algorithm development, we critically evaluate foundational signal processing practices, such as spike sorting, to promote analytic rigor and ensure reproducibility across the broader field of computational neuroscience.
