Figure 1. Pre-burst dynamics in STN recordings versus white noise. STN neural data are presented from 5 patients (10 hemispheres) with bilateral implants of the Summit RC+S neural interface (Medtronic), while white noise data are generated from 5 randomly selected seeds. (A–B) Z-score normalised, threshold-aligned burst envelopes for long (above median) and short (below median) duration bursts in white noise (A) and STN recordings (B). In white noise, long and short bursts exhibit similar pre-burst profiles, whereas in STN data prolonged bursts show a deeper dip and higher baseline amplitude. (C–D) Threshold-aligned envelopes after exclusion of bursts with inter-burst intervals < 400 ms, for white noise (C) and real data (D). The pre-burst dip is substantially attenuated, indicating that it is largely driven by preceding burst offsets. (E–F) Histograms of inter-burst intervals (E) and burst durations (F) for white noise and real data, demonstrating distinct temporal statistics. (G) Burst prediction performance is shown for the sliding window approach described in [3], using causally filtered broadband data (4-90 Hz) as input to the neural network. Five burst prediction performance metrics…
Key points are not available for this paper at this time.
No abstract yet. The journal hasn't published one we can show.
No takes yet. Share an insight, caveat, or question.
Abdi‐Sargezeh et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: