关键词: 孤独症谱系障碍, 瞳孔测量, 筛查诊断, 早期预测, 自主神经系统

Abstract: This review provides a systematic synthesis of empirical studies on the application of pupillometry in the auxiliary screening and diagnosis of children with autism spectrum disorder (ASD), establishing a unified analytical framework that encompasses three paradigms: the pupillary light reflex, complex visual stimulus processing, and attentional orienting. By leveraging multi-dimensional pupillary parameters, it enables objective differentiation between ASD and typically developing (TD) children, thereby addressing the limitations of traditional diagnostic tools that rely heavily on subjective questionnaires and often fail to capture early behavioral characteristics in young children.
First, this review stratifies and synthesizes the specific pupillary response patterns exhibited by ASD across different experimental paradigms, yielding a comprehensive set of physiological markers for early identification. In pupillary light reflex tasks, ASD consistently show prolonged constriction latency and reduced relative constriction amplitude, directly reflecting impaired parasympathetic regulation. Longitudinal evidence confirms that these physiological abnormalities are detectable in high-risk infants aged 6~24 months before the emergence of behavioral symptoms, with developmental stagnation of latency serving as a distinctive early risk indicator. When ASD are exposed to complex visual stimuli, the review identifies a stimulus-specific pupillary dissociation: phasic pupillary dilation triggered by social cues such as human faces and social scenes is markedly attenuated, whereas dilation to non-social stimuli—including geometric figures and objects related to circumscribed interests—is comparable to or even greater than that in typically developing children. This pattern reflects impaired automatic salience processing of social information in the brain. In exogenous attentional orienting tasks, ASD exhibit exaggerated pupillary dilation attributable to attentional disengagement deficits; in endogenous visual search tasks, they display elevated tonic baseline pupil size coupled with superior search efficiency relative to TD. The discriminative power of combined multi-task pupillary indices substantially exceeds that of single parameters; integrated pupillary light reflex indices alone achieve a classification accuracy of 81.5%.
Second, this review constructs a unified neurophysiological model linking pupillary abnormalities to core clinical symptoms of ASD, moving beyond the fragmented interpretations of single mechanisms prevalent in prior research. Three interconnected neural circuits are integrated to explain atypical pupillary responses. Deficient parasympathetic function and abnormal cerebellar regulation of the midbrain Edinger-Westphal nucleus delay pupillary light reflex conduction and reduce constriction amplitude, constituting the peripheral physiological basis of sensory atypicalities in ASD. Imbalanced tonic-phasic coupling within the locus coeruleus-norepinephrine (LC-NE) system disrupts the ventral frontoparietal attention network, leading to impaired salience processing of social stimuli and difficulties in attentional disengagement; at the same time, excessively enhanced phasic activity produces a pattern of uniform attentional allocation across the visual field, accounting for the visual search advantage observed in ASD. Dysregulated amygdala-prefrontal emotional circuitry gives rise to highly heterogeneous pupillary arousal in response to emotional faces, with this response pattern modulated by stimulus familiarity, gaze direction, and level of conscious processing. This three-tier mechanistic framework establishes correspondences between peripheral pupillary signals and central neural circuit abnormalities, providing objective physiological evidence for social communication impairments and atypical sensory processing in ASD.
Third, this review proposes a phased clinical translation roadmap for pupillometry and outlines standardized research protocols to address existing methodological bottlenecks. Two clinical application scenarios are delineated: for high-risk infants, pupillary light reflex parameters can be used for early risk stratification before behavioral symptoms emerge; for preschool-aged children with ambiguous behavioral presentations, pupillary responses elicited by social visual stimuli can complement existing behavioral assessments and shorten the diagnostic trajectory. To address key confounds—including inconsistent baseline collection standards, eye movement artifacts, and sample heterogeneity—priority directions for future research are proposed: developing standardized paradigms for baseline acquisition and data processing; conducting long-term longitudinal tracking of high-risk infant cohorts; constructing multimodal neural markers by integrating brain imaging and electrophysiological techniques; and performing cross-racial and cross-cultural validation to establish the generalizability of pupillary indices.
In conclusion, this review integrates previously scattered laboratory-based pupillary indices into a systematic and actionable objective screening framework for ASD. It identifies quantitative physiological biomarkers that precede behavioral symptoms, delineates the multi-level regulatory mechanisms underlying atypical autonomic, attentional, and emotional processing in ASD, and offers actionable specifications for clinical translation.

Key words: children with autism spectrum disorder, pupillometry, screening and diagnosis, early prediction, autonomic nervous system