CAJ | 학술논문

目的：探讨基于经验模式分解的心肺耦合技术在睡眠分析中的应用。方法通过分析30例源自临床多导睡眠监测记录的胸导心电信号，利用经验模式分解方法获得瞬时频率和瞬时相位信息，构建心肺耦合图谱。按照睡眠中的循环交替模式（CAP）分期方法，将睡眠过程划分为CAP、non-CAP以及清醒/异相睡眠期，采取过零率（ZCR）衡量心肺耦合最大峰值的波动程度，反映阻塞性睡眠呼吸暂停与低通气综合征（OSAHS）的严重程度。结果 OSAHS患者谱图分布频带集中于低频区域，各时刻最大峰值波动较小。通过对比人工和自动分期结果，可以发现基于经验模式分解的心肺耦合技术可以很精准地区分睡眠过程的不同状态。OSAHS患者和健康受试者在睡眠心肺耦合图谱中的最大峰值波动规律存在显著差异，健康组和轻/中度组的ZCR值具有显著性差异（P＜0.001）；轻/中度组与重度组的ZCR值存在显著性差异（P＜0.001），因此耦合最大峰值和睡眠呼吸暂停低通气指数均可以作为划分OSAHS严重程度的指标，且二者之间具有极强的负相关性（r=-0.77，P=5.8×10-18）。结论采用经验模式分解结合心肺耦合技术的方法，可提供可靠的睡眠微结构以及睡眠呼吸障碍信息，其数据采集简单、容易执行，在可穿戴健康管理以及临床辅助诊断领域有巨大发展潜力。
목적：탐토기우경험모식분해적심폐우합기술재수면분석중적응용。방법통과분석30례원자림상다도수면감측기록적흉도심전신호，이용경험모식분해방법획득순시빈솔화순시상위신식，구건심폐우합도보。안조수면중적순배교체모식（CAP）분기방법，장수면과정화분위CAP、non-CAP이급청성/이상수면기，채취과령솔（ZCR）형량심폐우합최대봉치적파동정도，반영조새성수면호흡잠정여저통기종합정（OSAHS）적엄중정도。결과 OSAHS환자보도분포빈대집중우저빈구역，각시각최대봉치파동교소。통과대비인공화자동분기결과，가이발현기우경험모식분해적심폐우합기술가이흔정준지구분수면과정적불동상태。OSAHS환자화건강수시자재수면심폐우합도보중적최대봉치파동규률존재현저차이，건강조화경/중도조적ZCR치구유현저성차이（P＜0.001）；경/중도조여중도조적ZCR치존재현저성차이（P＜0.001），인차우합최대봉치화수면호흡잠정저통기지수균가이작위화분OSAHS엄중정도적지표，차이자지간구유겁강적부상관성（r=-0.77，P=5.8×10-18）。결론채용경험모식분해결합심폐우합기술적방법，가제공가고적수면미결구이급수면호흡장애신식，기수거채집간단、용역집행，재가천대건강관리이급림상보조진단영역유거대발전잠력。
Objective To explore application of the EMD (Empirical Mode Decomposition)-based CPC (Cardio-Pulmonary Coupling) technique in sleep analysis. Methods Through analysis of 30 cases of thoracic ECG signals recorded by PSG (Poly-Somno-Graphy), the instantaneous frequency and instantaneous phase were obtained with deployment of EMD so as to construct a CPC map. Then, CAP (Cyclic Alternating Pattern) was utilized to divide sleep into three stages: CAP Stage, Non-CAP Stage and Wake/REM (Rapid Eyes Movement) Stage. The waving degree of the maximum CPC peak was measured by ZCR (Zero Crossing Rate), which reflected the severity of OSAHS (Obstructive Sleep Apnea-Hypopnea Syndrome). Results The frequency band of OSAHS patients’ map was distributed centralizedly in the low-frequency areas with small waving changes of the maximum peak at each time. Comparisons were made between manual staging and automatic staging, which revealed that EMD-based CPC could differentiate accurately between the different sleep statuses. Significant differences existed between the waving principles of the maximum peak in OSAHS Patient Group and Healthy Volunteer Group. ZCR values were significantly different between Slight/Middle OSAHS Patient Group and Healthy Volunteer Group (P<0.001), and between Slight/Middle OSAHS Patient Group and Severe OSAHS Patient Group (P<0.001). Therefore, the maximum coupling peak value and apnea-hypopnea Index could be used as indexes to identify the different severity of OSAHS patients. Moreover, strong negative correlation was seen between the two indexes (r=-0.77,P=5.8×10-18).Conclusion Combination of EMD and the CPC technique had proven its easy-to-operate features in data acquisition so as to provide reliable micro-structure and disorder information of sleep, which had huge development potentials in the ifelds of wearable health management and clinically-aided diagnosis.