Computer Vision for Brain Disorders Based Primarily on Ocular Responses

A modular machine learning tool for holistic and fine-grained behavioral analysis

Artificial intelligence techniques offer promising avenues for exploring human body features from videos, yet no freely accessible tool has reliably provided holistic and fine-grained behavioral analyses to date. To address this, we developed a machine learning tool based on a two-level approach: a first lower-level processing using computer vision for extracting fine-grained and comprehensive behavioral features such as skeleton or facial points, gaze, and action units; a second level of machine learning classification coupled with explainability providing modularity, to determine which behavioral features are triggered by specific environments. To validate our tool, we filmed 16 participants across six conditions, varying according to the presence of a person ("Pers"), a sound ("Snd"), or silence ("Rest"), and according to emotional levels using self-referential ("Self") and control ("Ctrl") stimuli. We demonstrated the effectiveness of our approach by extracting and correcting behavior from videos using two computer vision software (OpenPose and OpenFace) and by training two algorithms (XGBoost and long short-term memory [LSTM]) to differentiate between experimental conditions. High classification rates were achieved for "Pers" conditions versus "Snd" or "Rest" (AUC = 0.8-0.9), with explainability revealing actions units and gaze as key features. Additionally, moderate classification rates were attained for "Snd" versus "Rest" (AUC = 0.7), attributed to action units, limbs and head points, as well as for "Self" versus "Ctrl" (AUC = 0.7-0.8), due to facial points. These findings were consistent with a more conventional hypothesis-driven approach. Overall, our study suggests that our tool is well suited for holistic and fine-grained behavioral analysis and offers modularity for extension into more complex naturalistic environments.

Autosomal recessive cerebellar ataxias: a diagnostic classification approach according to ocular features

Autosomal recessive cerebellar ataxias (ARCAs) are a heterogeneous group of neurodegenerative disorders affecting primarily the cerebellum and/or its afferent tracts, often accompanied by damage of other neurological or extra-neurological systems. Due to the overlap of clinical presentation among ARCAs and the variety of hereditary, acquired, and reversible etiologies that can determine cerebellar dysfunction, the differential diagnosis is challenging, but also urgent considering the ongoing development of promising target therapies. The examination of afferent and efferent visual system may provide neurophysiological and structural information related to cerebellar dysfunction and neurodegeneration thus allowing a possible diagnostic classification approach according to ocular features. While optic coherence tomography (OCT) is applied for the parametrization of the optic nerve and macular area, the eye movements analysis relies on a wide range of eye-tracker devices and the application of machine-learning techniques. We discuss the results of clinical and eye-tracking oculomotor examination, the OCT findings and some advancing of computer science in ARCAs thus providing evidence sustaining the identification of robust eye parameters as possible markers of ARCAs.

Retina as a potential biomarker in schizophrenia spectrum disorders: a systematic review and meta-analysis of optical coherence tomography and electroretinography

INTRODUCTION

Abnormal findings on optical coherence tomography (OCT) and electroretinography (ERG) have been reported in participants with schizophrenia spectrum disorders (SSDs). This study aims to reveal the pooled standard mean difference (SMD) in retinal parameters on OCT and ERG among participants with SSDs and healthy controls and their association with demographic characteristics, clinical symptoms, smoking, diabetes mellitus, and hypertension.

METHODS

Using PubMed, Scopus, Web of Science, and PSYNDEX, we searched the literature from inception to March 31, 2023, using specific search terms. This study was registered with PROSPERO (CRD4202235795) and conducted according to PRISMA 2020.

RESULTS

We included 65 studies in the systematic review and 44 in the meta-analysis. Participants with SSDs showed thinning of the peripapillary retinal nerve fiber layer (pRNFL), macular ganglion cell layer- inner plexiform cell layer, and retinal thickness in all other segments of the macula. A meta-analysis of studies that excluded SSD participants with diabetes and hypertension showed no change in results, except for pRNFL inferior and nasal thickness. Furthermore, a significant difference was found in the pooled SMD of pRNFL temporal thickness between the left and right eyes. Meta-regression analysis revealed an association between retinal thinning and duration of illness, positive and negative symptoms. In OCT angiography, no differences were found in the foveal avascular zone and superficial layer foveal vessel density between SSD participants and controls. In flash ERG, the meta-analysis showed reduced amplitude of both a- and b-waves under photopic and scotopic conditions in SSD participants. Furthermore, the latency of photopic a-wave was significantly shorter in SSD participants in comparison with HCs.

DISCUSSION

Considering the prior report of retinal thinning in unaffected first-degree relatives and the results of the meta-analysis, the findings suggest that retinal changes in SSDs have both trait and state aspects. Future longitudinal multimodal retinal imaging studies are needed to clarify the pathophysiological mechanisms of these changes and to clarify their utility in individual patient monitoring efforts.

Artificial intelligence in dementia

PURPOSE OF REVIEW

Artificial intelligence tools are being rapidly integrated into clinical environments and may soon be incorporated into dementia diagnostic paradigms. A comprehensive review of emerging trends will allow physicians and other healthcare providers to better anticipate and understand these powerful tools.

RECENT FINDINGS

Machine learning models that utilize cerebral biomarkers are demonstrably effective for dementia identification and prediction; however, cerebral biomarkers are relatively expensive and not widely available. As eye images harbor several ophthalmic biomarkers that mirror the state of the brain and can be clinically observed with routine imaging, eye-based machine learning models are an emerging area, with efficacy comparable with cerebral-based machine learning models. Emerging machine learning architectures like recurrent, convolutional, and partially pretrained neural networks have proven to be promising frontiers for feature extraction and classification with ocular biomarkers.

SUMMARY

Machine learning models that can accurately distinguish those with symptomatic Alzheimer's dementia from those with mild cognitive impairment and normal cognition as well as predict progressive disease using relatively inexpensive and accessible ocular imaging inputs are impactful tools for the diagnosis and risk stratification of Alzheimer's dementia continuum. If these machine learning models can be incorporated into clinical care, they may simplify diagnostic efforts. Recent advancements in ocular-based machine learning efforts are promising steps forward.