Segmentation Procedure for Spontaneous Local Field Potential
Open AccessLocal field potential (LFP) is an electrophysiological signal that reflects summed electrical activity of nearby neurons. Temporal segmental structure has be found in spontaneous LFP, which result from the brain switching between global states. Recognition of the states, or in other word, segmentation of the signal, serves as an indispensible reference in neuroscience research. Major obstacles of existing approaches are reviewed first, then a segmentation approach is suggested based on another perspective of the problem. After explanation of mathematical tools that will be used, an entire segmentation procedure for multi-channel spontaneous LFP measured from rat visual cortex, is illustrated step by step, including necessary preprocessing. To start with, preparation of data is described, then further preprocessing to acquire LFP signal for segmentation is elaborated, including locating intracranial channels based on inter-channel correlation coefficients, band-pass filtering without distortion, and selecting a single channel in anatomical layer of interest through principle component analysis (PCA). To better reveal segmental feature, the signal is further represented in time-frequency domain through multi-taper power spectrum density (PSD) estimation. Based on PSD, the signal is segmented through non-negative matrix factorization (NMF) with enforced sparseness on coefficients. Results from both simulated data and real data are provided to show the effectiveness of the procedure. Promising aspects of the procedure and future work are summarized in the end.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.