Machine Learning Algorithm for Recognition of Neurological Disorders Using a Multichannel Battery-free Wireless Brain Implant Recorder
Conference
Gutierrez-Hernandez, M, Moncion, C, Bojja-Venkatakrishnan, S et al. (2023). Machine Learning Algorithm for Recognition of Neurological Disorders Using a Multichannel Battery-free Wireless Brain Implant Recorder
. 258-261. 10.1109/EMTS57498.2023.10925312
Gutierrez-Hernandez, M, Moncion, C, Bojja-Venkatakrishnan, S et al. (2023). Machine Learning Algorithm for Recognition of Neurological Disorders Using a Multichannel Battery-free Wireless Brain Implant Recorder
. 258-261. 10.1109/EMTS57498.2023.10925312
Automatic monitoring of neural activity can lead to the recognition of neurological patient disorders in real time. In this study, we present an evaluation of a multichannel, battery-free, and fully implantable wireless neurosensing system (WiNS). Signal quality from WiNS is compared with a commercially available wired system employing machine learning algorithms. Somatosensory evoked potential from fore limb and Whisker stimulation paradigms are considered. Machine learning testing accuracy from wired and our new wireless system shows < 8% difference in Whisker and fore limb stimulations. Our wireless system also indicates successful detection of biopotential signals as low as 15 μVpp. These findings support the premise that the proposed real-time machine interface is effective for wireless and battery-free acquisition of biopotentials. Given this development, system evaluation and classification can be pursued for future clinical research.
