Decoding Brain Signals in a Neuromorphic Framework for a Personalized Adaptive Control of Human Prosthetics DOI Creative Commons

G. Rusev,

Svetlozar Yordanov,

Simona Nedelcheva

et al.

Biomimetics, Journal Year: 2025, Volume and Issue: 10(3), P. 183 - 183

Published: March 14, 2025

Current technological solutions for Brain-machine Interfaces (BMI) achieve reasonable accuracy, but most systems are large in size, power consuming and not auto-adaptive. This work addresses the question whether current neuromorphic technologies could resolve these problems? The paper proposes a novel framework of BMI system prosthetics control via decoding Electro Cortico-Graphic (ECoG) brain signals. It includes three-dimensional spike timing neural network (3D-SNN) signals features extraction an on-line trainable recurrent reservoir structure (Echo state (ESN)) Motor Control Decoding (MCD). A software system, written Python using NEST Simulator SNN library is described. able to adapt continuously real time supervised or unsupervised mode. proposed approach was tested on several experimental data sets acquired from tetraplegic person. First simulation results encouraging, showing also need further improvement multiple hyper-parameters tuning. Its future implementation hardware platform that smaller size significantly less discussed too.

Language: Английский

Decoding Brain Signals in a Neuromorphic Framework for a Personalized Adaptive Control of Human Prosthetics DOI Creative Commons

G. Rusev,

Svetlozar Yordanov,

Simona Nedelcheva

et al.

Biomimetics, Journal Year: 2025, Volume and Issue: 10(3), P. 183 - 183

Published: March 14, 2025

Current technological solutions for Brain-machine Interfaces (BMI) achieve reasonable accuracy, but most systems are large in size, power consuming and not auto-adaptive. This work addresses the question whether current neuromorphic technologies could resolve these problems? The paper proposes a novel framework of BMI system prosthetics control via decoding Electro Cortico-Graphic (ECoG) brain signals. It includes three-dimensional spike timing neural network (3D-SNN) signals features extraction an on-line trainable recurrent reservoir structure (Echo state (ESN)) Motor Control Decoding (MCD). A software system, written Python using NEST Simulator SNN library is described. able to adapt continuously real time supervised or unsupervised mode. proposed approach was tested on several experimental data sets acquired from tetraplegic person. First simulation results encouraging, showing also need further improvement multiple hyper-parameters tuning. Its future implementation hardware platform that smaller size significantly less discussed too.

Language: Английский

Citations

0