作者: Brooks A. Gross , Christine M. Walsh , Apurva A. Turakhia , Victoria Booth , George A. Mashour
DOI: 10.1016/J.JNEUMETH.2009.07.009
关键词:
摘要: Abstract Manual state scoring of physiological recordings in sleep studies is time-consuming, resulting a data backlog, research delays and increased personnel costs. We developed MATLAB-based software to automate sleep/waking states rats, potentially extendable other animals, from variety recording systems. The contains two programs, Sleep Scorer Auto-Scorer, for manual automated scoring. Auto-Scorer logic-based program that displays power spectral densities an electromyographic (EMG) signal σ, δ, θ frequency bands electroencephalographic (EEG) signal, along with the δ/θ ratio σ × θ, every epoch. user defines thresholds training file definitions which uses logic discriminate epoch file. was evaluated by comparing its output manually scored files 6 rats under 2 experimental conditions 3 users. Each generated file, set thresholds, auto-scored 12 into 4 (waking, non-REM, transition-to-REM, REM sleep) 1/4 time required score Overall performance comparisons between resulted mean agreement 80.24 ± 7.87%, comparable average among scorers (83.03 ± 4.00%). There no significant difference user–user user–Auto-Scorer ratios. These results support use our open-source coupled review, rapidly accurately rat recordings.