Deep learning of spontaneous arousal fluctuations detects early cholinergic defects across neurodevelopmental mouse models and patients

Abstract

Neurodevelopmental spectrum disorders like autism (ASD) are diagnosed, on average, beyond age 4 y, after multiple critical periods of brain development close and behavioral intervention becomes less effective. This raises the urgent need for quantitative, noninvasive, and translational biomarkers for their early detection and tracking. We found that both idiopathic (BTBR) and genetic (CDKL5- and MeCP2-deficient) mouse models of ASD display an early, impaired cholinergic neuromodulation as reflected in altered spontaneous pupil fluctuations. Abnormalities were already present before the onset of symptoms and were rescued by the selective expression of MeCP2 in cholinergic circuits. Hence, we trained a neural network (ConvNetACh) to recognize, with 97% accuracy, patterns of these arousal fluctuations in mice with enhanced cholinergic sensitivity (LYNX1-deficient). ConvNetACh then successfully detected impairments in all ASD mouse models tested except in MeCP2-rescued mice. By retraining only the last layers of ConvNetACh with heart rate variation data (a similar proxy of arousal) directly from Rett syndrome patients, we generated ConvNetPatients, a neural network capable of distinguishing them from typically developing subjects. Even with small cohorts of rare patients, our approach exhibited significant accuracy before (80% in the first and second year of life) and into regression (88% in stage III patients). Thus, transfer learning across species and modalities establishes spontaneous arousal fluctuations combined with deep learning as a robust noninvasive, quantitative, and sensitive translational biomarker for the rapid and early detection of neurodevelopmental disorders before major symptom onset.

Genes and environments, development and time

“…Artoni et al (72) develop an approach for early detection of neurodevelopmental spectrum disorders, in this case, autism spectrum disorder (ASD), by using a transfer learning experiment across species (mouse to human). Their approach is based on spontaneous arousal fluctuations combined with deep learning and is a possible breakthrough in the early detection of risk for ASD and related disorders, where late diagnosis strongly diminishes intervention efficacy.…”

Evoked Potentials and EEG Analysis in Rett Syndrome and Related Developmental Encephalopathies: Towards a Biomarker for Translational Research

“…This includes brain-based measures, including magnetoencephalography or transcranial magnetic stimulation, pupillometry, and sympathetic testing, as well as physiological/behavioral measures derived from wearable sensors. Although few studies to date have utilized these methods in participants with Rett syndrome (Heinen and Korinthenberg, 1996;Heinen et al, 1997;Krajnc and Zidar, 2016;Santosh et al, 2017;Artoni et al, 2019), these approaches have proven useful in biomarker research for other neurodevelopmental disorders (Roberts et al, 2010;Oberman et al, 2016;Ness et al, 2017).…”

Modulation of cholinergic activity through lynx prototoxins: Implications for cognition and anxiety regulation

“…Although the role of lynx1 on disease states has focused primarily on Alzheimer’s, recently ( Artoni et al, 2019 ) utilized the enhanced cholinergic tone of lynx1 KO mice to investigate if alterations in cholinergic circuit alter arousal dynamics similar to those observed in mouse models of Autism Spectrum Disorder (ASM). These studies determined lynx1 KO mice exhibited a shifted distribution toward maximal pupil size similar to ASD model mice ( Artoni et al, 2019 ). lynx1 KO mice were also used to screen for genes differentially regulated and linked to genes in patients with risk for neurodevelopmental disorders such as epilepsy and schizophrenia ( Smith et al, 2018 ).…”