ICM (INSTITUT DU CERVEAU ET DE LA MOELLE EPINIERE);APHP (ASSISTANCE PUBLIQUE - HOPITAUX DE PARIS);CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE;INSERM (INSTITUT NATIONAL;BIOSERENITY;SORBONNE UNIVERSITE
The present invention relates to a computer-implemented method for detecting pathological brain activity patterns from a scalp electroencephalographic signal, the method comprising the steps of obtaining (A) an electroencephalographic signal as a function of multiple channels and time； identifying (C), for each channel, the zero-crossings of the electroencephalographic signal over a fixed threshold； generating a zero-crossing representation of at least a segment of the obtained electroencephalographic signal with the identified zero-crossings； obtaining (D) a reference family of real functions of time and channels from a zero-crossing statistical analysis of zero-crossing representation of pre-recorded electroencephalographic signals； calculating (E) a matching score by comparing said zero-crossing representation of a segment of the electroencephalographic signal with at least one reference function from the reference family of functions； and computing the matching score as a function of time by sliding the at least one reference function from the reference family of functions over the electroencephalographic signal.本發明涉及一種用於從頭皮腦電圖信號中檢測病理性腦活動模式的電腦實現的方法，該方法包括以下步驟：獲得（A）作為多個通道和時間的函數的腦電圖信號；為每個通道識別（C）在固定閾值上的腦電圖信號的過零點；利用所識別的過零點產生所獲得的腦電圖信號的至少一片段的過零點表示；從預先記錄的腦電圖信號的過零點表示的過零點統計分析中獲得（D）時間和通道的實函數的參考族；藉由比較所述腦電圖信號的片段的所述過零點表示與來自所述函數參考族的至少一個參考函數來計算（E）匹配評分；和藉由在腦電圖信號上滑動來自函數參考族的至少一個參考函數來計算作為時間的函數的匹配評分。
