CAJ | 학술논문

摘心理负荷的评估正经历着从传统工效学向神经人因学的变革。EEG、ERPs、fMRI、fNIRS和TCD等神经科学研究技术为这场变革提供了有力的工具。研究发现：在单任务情境下，随着操作者心理负荷的增加，脑电α波活动减弱，θ波活动增强，并且前额皮层血流、左侧额下回血液氧合血红蛋白浓度变化均增加，大脑动脉血流速度也增快：在双任务情境下，随着操作者主任务心理负荷的增加，次任务的N1、新异P3和P3b等ERPs成分波幅降低。依据这些研究成果，学者们利用人工神经网络、支持向量机等模式分类算法实现了对心理负荷的实时在线评估。但是，各种研究技术在敏感性、诊断力、主任务干扰、实施需求、可接受性和信度等方面各有优势与不足。在未来研究中，要注意促进相关技术的融合，提高其可接受性，并充分利用模式识别算法提高其诊断力和敏感性。
적심리부하적평고정경력착종전통공효학향신경인인학적변혁。EEG、ERPs、fMRI、fNIRS화TCD등신경과학연구기술위저장변혁제공료유력적공구。연구발현：재단임무정경하，수착조작자심리부하적증가，뇌전α파활동감약，θ파활동증강，병차전액피층혈류、좌측액하회혈액양합혈홍단백농도변화균증가，대뇌동맥혈류속도야증쾌：재쌍임무정경하，수착조작자주임무심리부하적증가，차임무적N1、신이P3화P3b등ERPs성분파폭강저。의거저사연구성과，학자문이용인공신경망락、지지향량궤등모식분류산법실현료대심리부하적실시재선평고。단시，각충연구기술재민감성、진단력、주임무간우、실시수구、가접수성화신도등방면각유우세여불족。재미래연구중，요주의촉진상관기술적융합，제고기가접수성，병충분이용모식식별산법제고기진단력화민감성。
The assessment of mental workload is undergoing transition from traditional ergonomics to neuroergonomics. The neuroimaging techniques such as electroencephalography （EEG）, event-related potentials （ERPs）, functional magnetic resonance imaging （fMRI）, functional Near-infrared spectroscopy （fNIRS） and transcranial doppler （TCD） provide the strong supports for this revolution. It is found that under single-task condition, with the increase of mental workload level, alpha-band power decreases, theta-band power increases. In addition, the cerebral blood flow in the regions of the prefrontal cortex （PFC）, the average oxygenation changes in the regions of left inferior frontal gyrus （LIFG） and the cerebral blood flow velocity （CBFV） also increase. Under dual-task condition, with the increase of mental workload level, the amplitudes of several ERPs components elicited by the secondary task stimuli （e.g., N1, Novelty P3, P3b, etc） decrease. Based on these findings, researchers have achieved the real-time online evaluation of mental workload by pattern classification algorithms （e.g.,artificial neural network, support vector machine）. But all of these techniques have their own advantages and disadvantages on several aspects （i.e., sensitivity, diagnosticity, primary task intrusion, implementation requirements, acceptability and reliability）. In the future, researchers should promote the combination of these neuroimaging technologies, improve their acceptability and enhance their sensitivity and diagnosticity via pattern recognition algorithms.