Deep Neural Networks and Brain Alignment: Brain Encoding and Decoding (Survey) DOI Creative Commons
Subba Reddy Oota, Manish Gupta, Raju S. Bapi

и другие.

arXiv (Cornell University), Год журнала: 2023, Номер unknown

Опубликована: Янв. 1, 2023

How does the brain represent different modes of information? Can we design a system that automatically understands what user is thinking? Such questions can be answered by studying recordings like functional magnetic resonance imaging (fMRI). As first step, neuroscience community has contributed several large cognitive datasets related to passive reading/listening/viewing concept words, narratives, pictures and movies. Encoding decoding models using these have also been proposed in past two decades. These serve as additional tools for basic research science neuroscience. aim at generating fMRI representations given stimulus automatically. They practical applications evaluating diagnosing neurological conditions thus help therapies damage. Decoding solve inverse problem reconstructing stimuli fMRI. are useful designing brain-machine or brain-computer interfaces. Inspired effectiveness deep learning natural language processing, computer vision, speech, recently neural encoding proposed. In this survey, will discuss popular language, vision speech stimuli, present summary datasets. Further, review based architectures note their benefits limitations. Finally, conclude with brief discussion about future trends. Given amount published work `computational neuroscience' community, believe survey nicely organizes plethora presents it coherent story.

Язык: Английский

Deep Neural Networks and Brain Alignment: Brain Encoding and Decoding (Survey) DOI Creative Commons
Subba Reddy Oota, Manish Gupta, Raju S. Bapi

и другие.

arXiv (Cornell University), Год журнала: 2023, Номер unknown

Опубликована: Янв. 1, 2023

How does the brain represent different modes of information? Can we design a system that automatically understands what user is thinking? Such questions can be answered by studying recordings like functional magnetic resonance imaging (fMRI). As first step, neuroscience community has contributed several large cognitive datasets related to passive reading/listening/viewing concept words, narratives, pictures and movies. Encoding decoding models using these have also been proposed in past two decades. These serve as additional tools for basic research science neuroscience. aim at generating fMRI representations given stimulus automatically. They practical applications evaluating diagnosing neurological conditions thus help therapies damage. Decoding solve inverse problem reconstructing stimuli fMRI. are useful designing brain-machine or brain-computer interfaces. Inspired effectiveness deep learning natural language processing, computer vision, speech, recently neural encoding proposed. In this survey, will discuss popular language, vision speech stimuli, present summary datasets. Further, review based architectures note their benefits limitations. Finally, conclude with brief discussion about future trends. Given amount published work `computational neuroscience' community, believe survey nicely organizes plethora presents it coherent story.

Язык: Английский

Процитировано

2