Back

Hardware-Powered Ultra Low Latency (HarPULL) Rhythmic State Dependent TMS Technology

Fedosov, N.; Makarova, M.; Nekrasova, J.; Ossadtchi, A.

2024-05-04 neurology
10.1101/2024.05.03.24306813 medRxiv
Show abstract

ObjectiveUpcoming neuroscientific research will require bidirectional and context dependent interaction with nervous tissue. To facilitate the future neuroscientific discoveries we have created HarPULL, a genuinely real-time system for tracking oscillatory brain state. ApproachThe HarPULL technology ensures reliable, accurate and affordable real-time phase and amplitude tracking based on the state-space estimation framework operationalized by Kalman filtering. To avoid data transfer delays and to obtain a truly real-time system the algorithm is implemented on the computational core of an EEG amplifier controlled by a real-time operating system. Systems performance is tested with simulated and real data both online and offline and within a real-time state dependent TMS using a phantom and human subjects. Main resultsWe show that taking into account the 1/f nature of the brain noise and the use of the steady state colored Kalman filter further improves phase tracking performance in both simulated and real data. We use HarPULL to trigger the TMS device contingent upon the target phase and amplitude combination and demonstrate minimal delay (2 ms) between the occurrence of the predetermined rhythm phase in the cortex and the corresponding magnetic stimulus. Using this setup in the real-time setting we observe a significant modulation of the motor evoked potentials (MEP) by the sensorimotor rhythms state. Finally, we use HarPULL and for the first time obtain phase-dependent muscle cortical representation (MCR) maps in real-time. We show better delineation between the representations of several muscles when the stimulation is performed in the excitation state. SignificanceHarPULL is the first truly real-time technology for the instantaneous tracking of the brains rhythmic activity. Our technological solution establishes a nearly instantaneous non-invasive contact with a living brain which has a broad range of clinical, diagnostic and scientific applications.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.