Research Topic

Large-Scale On-Chip Neural Computation Catalyzed by Machine Learning

About this Research Topic

The fast advancements in electrophysiology technologies on microelectrodes and electronics have allowed data acquisition from a large number of channels simultaneously, creating a tremendous and ever-increasing amount of data. The pace of technological change has introduced great computational challenges to emerging data analysis techniques that should scale well with the amount of data and advance the studies on how to process, transmit, and interpret such exploding data volumes to interact with physical world. There are several critical limitations of present data processing techniques that limit broadspread adaptation of current high-density electrophysiological technologies, including (i) transmitting large amount of data through limited communication bandwidth (e.g., wireless) with minimum information loss, (ii) achieving real-time and large-scale signal processing under stringent budget of hardware resources for implantable electronics, and (iii) interpreting electrophysiological signals to interact with complex physical world in a robust and precise fashion.

Recent years have seen a surge of machine learning (ML) applications in neural signal processing to address the aforementioned challenges, thanks to both the substantial theoretical progress and the availability of high-performance and massively-parallelable computational resources (e.g., graphics processing units), that jointly provide the capability to handle large volume of neural data in both online and offline modes. To meet the real-time and scalability requirements, optimizing ML algorithms and turning them into on-chip units pose new challenges in balancing algorithmic precision, versatility, and implementability in hardware. Successful design and implementation of such ML capabilities necessitate solid understanding of both characteristics of neural data being recorded and algorithms, in particularly from a hardware and system perspective.

Considering these facts, our Research Topic aims to collect contributions in all article types (reviews, research papers, methodology papers, etc.) that are broadly related to on-chip design of ML algorithms and its applications to neurophysiological experiments, including neural recording, neural stimulation, nerve decoding, prostheses control, and many more. We welcome submissions of research articles that leverage advanced ML developments to uncover new patterns or insights from neural signals that strengthens our understanding and interaction of nervous system.

Specific topic may include, but not limited to:
- ML methods that reduce data bandwidth and enable wireless transmission, such as autoencoder, compressive sensing, etc.
- Spike sorting algorithms fueled by recent ML developments
- Neural decoding algorithms based on ML approaches that extract motor related signals from EEG/ENG/ECoG/EMG to restore motion functions
- Neuromorphic engineering
- ML based algorithms that enhance signal quality of neural recordings
- Efficient and scalable hardware implementation of ML based neural signal processing algorithms.


Keywords: neural signal processing, machine learning, on-chip computation, neural instrumentation, hardware-algorithm optimization


Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.

The fast advancements in electrophysiology technologies on microelectrodes and electronics have allowed data acquisition from a large number of channels simultaneously, creating a tremendous and ever-increasing amount of data. The pace of technological change has introduced great computational challenges to emerging data analysis techniques that should scale well with the amount of data and advance the studies on how to process, transmit, and interpret such exploding data volumes to interact with physical world. There are several critical limitations of present data processing techniques that limit broadspread adaptation of current high-density electrophysiological technologies, including (i) transmitting large amount of data through limited communication bandwidth (e.g., wireless) with minimum information loss, (ii) achieving real-time and large-scale signal processing under stringent budget of hardware resources for implantable electronics, and (iii) interpreting electrophysiological signals to interact with complex physical world in a robust and precise fashion.

Recent years have seen a surge of machine learning (ML) applications in neural signal processing to address the aforementioned challenges, thanks to both the substantial theoretical progress and the availability of high-performance and massively-parallelable computational resources (e.g., graphics processing units), that jointly provide the capability to handle large volume of neural data in both online and offline modes. To meet the real-time and scalability requirements, optimizing ML algorithms and turning them into on-chip units pose new challenges in balancing algorithmic precision, versatility, and implementability in hardware. Successful design and implementation of such ML capabilities necessitate solid understanding of both characteristics of neural data being recorded and algorithms, in particularly from a hardware and system perspective.

Considering these facts, our Research Topic aims to collect contributions in all article types (reviews, research papers, methodology papers, etc.) that are broadly related to on-chip design of ML algorithms and its applications to neurophysiological experiments, including neural recording, neural stimulation, nerve decoding, prostheses control, and many more. We welcome submissions of research articles that leverage advanced ML developments to uncover new patterns or insights from neural signals that strengthens our understanding and interaction of nervous system.

Specific topic may include, but not limited to:
- ML methods that reduce data bandwidth and enable wireless transmission, such as autoencoder, compressive sensing, etc.
- Spike sorting algorithms fueled by recent ML developments
- Neural decoding algorithms based on ML approaches that extract motor related signals from EEG/ENG/ECoG/EMG to restore motion functions
- Neuromorphic engineering
- ML based algorithms that enhance signal quality of neural recordings
- Efficient and scalable hardware implementation of ML based neural signal processing algorithms.


Keywords: neural signal processing, machine learning, on-chip computation, neural instrumentation, hardware-algorithm optimization


Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.

About Frontiers Research Topics

With their unique mixes of varied contributions from Original Research to Review Articles, Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author.

Topic Editors

Loading..

Submission Deadlines

30 September 2021 Abstract
12 December 2021 Manuscript

Participating Journals

Manuscripts can be submitted to this Research Topic via the following journals:

Loading..

Topic Editors

Loading..

Submission Deadlines

30 September 2021 Abstract
12 December 2021 Manuscript

Participating Journals

Manuscripts can be submitted to this Research Topic via the following journals:

Loading..
Loading..

total views article views article downloads topic views

}
 
Top countries
Top referring sites
Loading..