Characterization of Neural Interaction During Learning and Adaptation from Spike-Train Data
A basic task in understanding the neural mechanism of learning andadaptation is to detect and characterize neural interactions andtheir changes in response to new experiences. Recent experimentalwork has indicated that neural interactions in the primary motorcortex of the monkey brain tend to change...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2004-10-01
|
| Series: | Mathematical Biosciences and Engineering |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/doi/10.3934/mbe.2005.2.1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A basic task in understanding the neural mechanism of learning andadaptation is to detect and characterize neural interactions andtheir changes in response to new experiences. Recent experimentalwork has indicated that neural interactions in the primary motorcortex of the monkey brain tend to change their preferreddirections during adaptation to an external force field. Toquantify such changes, it is necessary to develop computationalmethodology for data analysis. Given that typical experimentaldata consist of spike trains recorded from individual neurons,probing the strength of neural interactions and their changes isextremely challenging. We recently reported in a briefcommunication [Zhu et al., Neural Computations 15 ,2359 (2003)] a general procedure to detect and quantify the causalinteractions among neurons, which is based on the method ofdirected transfer function derived from a class of multivariate,linear stochastic models. The procedure was applied to spiketrains from neurons in the primary motor cortex of the monkeybrain during adaptation, where monkeys were trained to learn a newskill by moving their arms to reach a target under externalperturbations. Our computation and analysis indicated that theadaptation tends to alter the connection topology of theunderlying neural network, yet the average interaction strength inthe network is approximately conserved before and after theadaptation. The present paper gives a detailed account of thisprocedure and its applicability to spike-train data in terms ofthe hypotheses, theory, computational methods, control test, andextensive analysis of experimental data. |
|---|---|
| ISSN: | 1551-0018 |