Japanese
English
- 有料閲覧
- Abstract 文献概要
- 1ページ目 Look Inside
- 参考文献 Reference
はじめに
ブレイン・マシン・インターフェース(brain machine interface:BMI),あるいはブレイン・コンピュータ・インターフェース(brain computer interface:BCI)とは,脳の活動を計測して機械やコンピュータに送り,そこから有用な情報を読み出して使用する技術である。その応用の取り組みは,四肢などの運動器に損傷がある患者の脳や脊髄から運動意図に関連した神経信号を読み出し車椅子やロボットアームなどの機械を動かす,「運動出力型BMI」で最も盛んに行われている。しかし,筋萎縮性側索硬化症(amyotrophic lateral sclerosis:ALS)や「閉じ込め症候群」などさらに重度の運動麻痺を持つ患者は,介護者との日々の意思疎通にすら困難をきたすケースが少なくない。
筆者らは,こうした重症患者が視野の中で注意を向けた対象物の情報や,頭に思い浮かべたイメージの情報などを読み取り,他者(例えば介護者)に伝えるBMI/BCIの開発を研究の最終目標としている(Fig.1)。本稿ではその第一段階として筆者らが行った,神経の電気的活動を低侵襲的に記録できる皮質脳波記録法(electrocorticogram:ECoG)を用いて視覚入力のカテゴリー情報を読み出す研究を紹介する。
Abstract
Electrocorticogram (ECoG) is an electrophysiological brain activity recording technique that has been widely revisited in recent years, not only for clinical monitoring, but also for prosthetic applications. However, the extent and limitations of the technique are poorly understood. Higher areas of human and macaque ventral visual cortices are known to have functional domain structures that are selective to certain categories, and population vectors that have been derived from visually evoked single-unit activity (SUA) recording in this region have been shown to form category clusters. How can visually evoked potentials recorded with ECoG from the same region be exploited to extract category information? To answer this question, the development of a simultaneous ECoG and SUA recording device by the modification of a previously reported flexible mesh ECoG probe with a microelectromechanical system has been promising (Toda et al., 2011). Indeed, Toda et al. conducted simultaneous recordings and reported that mesh ECoG signals exhibited comparable or better signal variabilities compared to conventional methods in the rat visual cortex. With this approach, we conducted intensive simultaneous ECoG and SUA recordings from the macaque anterior inferior temporal (IT) cortex. We compared how basic visual category and fine information is decoded from different recording modalities. Our preliminary results indicated that ECoG signals from the IT cortex may be a useful source for reading out certain levels of category information from visual input.
Copyright © 2013, Igaku-Shoin Ltd. All rights reserved.