画像認知の根源的課題 : 脳内ハードウエア(一般セッション3 画像認識,その他,計算解剖モデルに基づく診断・治療支援,医用画像処理一般)

概要

論文の詳細を見る
Neuron is a simple computational element with rough function, e.g. it cannot receive spikes between refractory period of after it fires. Besides this, its parameters fluctuate time to time. However, as a neuron network they can work intellectually. This gap is an big problem to be solved not only from physiological point of view but also from computer science. We detected pseudo random sequences including M-sequence family in spike trains of cultured neuron network significantly more than by chance. This gave code flow map on multi-electrode array dishes (MED-P515A, Alpha MED Scientific, Kadoma, Osaka, Japan) with 8 × 8 planar microelectrodes. Each electrode catches spikes from several neurons. Conventionally these spike trains are sorted to extract spikes from respective single neuron. Unfortunately but fortunately, however, we obtained mixed neuron spike trains, and by analyzing these, we could find principle of intelligent neuronal network processing.
2014-01-19