心房顫動是因心房內非正常起搏點位置出現不正常快速電氣活動,掩蓋原本節律所產生。而一個世紀以來,對於心房顫動是由深藏在純粹隨機性表象下的少數特定源頭驅動,抑或僅是碎裂波不規則隨機碰撞之產物仍未有定論。主張存在穩定轉子或脈波擴散源作為「驅動器」並維持心房顫動,對之進行電氣燒灼已證實可改善手術結果。然「驅動器」偵測自動化目前仍有瓶頸尚待克服,如偵測導管導極數目有限,覆蓋範圍不足、驅動源附近易產生碎裂波干擾、及「驅動器」可能出現空間中小幅漫移與間歇性激發等,均使持續型AF的局部驅動源難以偵測。本團隊已開發可兼容多種導管的前級電生理訊號放大器,配合三維標測系統的指標圖像化、後端工作站數據分析,可對複雜個案進行高密度標測以分析即時電氣傳導路徑,並利用稀疏學習增強AF「驅動器」預測,判定需處理的位置。本計畫希望證實隱沒於周遭高度碎裂波下,真正維持AF存在處,可透過即時高密度空間採樣訊號,標測於新的心房基質影像。本計畫亦將探討於竇性節律下,利用高密度心房基質標測(時頻分析指標)尋找重度纖維化區域之可行性。期望最終以新策略協助提升持續型及長期持續型AF病患治癒率,大幅縮短定位及電燒時間並減小電燒面積。 ;Atrial fibrillation (AF) may not always be triggered by a fully random process. The stable and rapid re-entrant circuits resulting in fibrillatory conduction throughout the atria can persist for minutes even hours. Ablation at the center of stable rotating waves and focal sources resulted in a high rate of acute AF termination and a better long-term recurrence-free probability. However, during mapping the atrial substrate electrograms of AF, a frequently encountered difficulty is the identification of culprit sites and the analysis of the wave propagation particularly for the electrogram signals with wide temporal and spatial disparities. Localizing AF drivers by the conventional sequential temporal-spatial mapping in the persistent AF is even harder because of its lack of specificity of complex atrial electrograms, intermittent firing, and spatial meandering. Our newly developed system with the 64 channels amplifier frontend can be compatible with different types of catheters and achieve optimal computing efficiency in real-time ultra-high density mapping of the atrial substrates by implementing a heterogeneous computation. The system can be used to interpret the complicated wave propagation and identify the substrate that maintains AF through the multi-task sparse learning, which helps to locate the true AF driver hidden beneath the highly fragmented waves. In this project, to improve the efficiency of procedure, we also propose a novel strategy of SAFE-T (Simultaneous Amplitude Frequency Electrogram Transformation) and sinus rhythm (SR)- abnormal electrogram-guided ablation. We hypothesize that the novel real-time temporal frequency analysis can be used to identify and characterize the abnormal substrates even during SR, and the removal of potential redundant conducting channels in persistent AF can assure the long-term SR maintenance. In conclusion, we aim to develop the automated electrogram analysis using real-time ultra-high density substrate mapping which allows instantaneous and objective identification of abnormal potential that accurately indicates AF driver. Our new strategy aims not only to reduce the ablation area but also to improve the acute termination rate and the recurrence-free survival of AF.