Survivorship Bias
研究发现，脑电图中的非周期性背景活动（频谱斜率）比传统频带功率更能准确区分精神分裂症患者与健康对照，且优于患者自身行为表现，可作为更可靠的生物标志物。
<h3>Abstract</h3> Diagnosis and symptom severity in schizophrenia is associated with irregularities across neural oscillatory frequency bands, including theta, alpha, beta, and gamma. However, electroencephalographic signals consist of both periodic and aperiodic background activity, characterized by the (1/f) slope of the power spectrum. We compared traditional band-limited periodic oscillatory activity to aperiodic activity, in schizophrenia participants and healthy controls, during an attention task. Classification analysis revealed that the change in slope of the power spectrum better predicted group status than traditional oscillatory power. It even outperformed the predictions made using participants’ own behavioral performance. Additionally, the differences in slope were highly consistent as they were observed across all electrodes. In sum, the aperiodic slope appears to be a more accurate, consistent, and robust metric to differentiate schizophrenic patients from healthy controls. <h3>Significance statement</h3> Understanding the neurobiological origins of schizophrenia and identifying reliable and consistent biomarkers are of critical importance in improving treatment of that disease. Numerous studies have reported disruptions to neural oscillations in schizophrenia patients. This has, in part, led to schizophrenia being characterized as a disease of disrupted neural coordination, reflected by changes in frequency band power. We report however that changes in aperiodic background noise (i.e., spectral slope) can also predict clinical status. Unlike band-limited power though, the aperiodic slope predicts status better than participants’ own behavioral performance. Furthermore, it is a consistent predictor across all electrodes. Alterations in aperiodic noise are consistent with well-established inhibitory neuron dysfunctions associated with schizophrenia, allowing for a direct link between noninvasive EEG and chronic, widespread, neurobiological deficits.