Table 1 Emerging predictive technologies. Table includes examples of completed and ongoing prospective studies using biomarker technologies to capture biological or behavioral differences in the first years of life that predict subsequent autism diagnosis. PPV = positive predictive value; NPV = negative predictive value
From: Parent attitudes towards predictive testing for autism in the first year of life
Lead Author/Contact PI | Year published/awarded | Journal/Grant# | Method | Population | Age | Sensitivity | Specificity | PPV | NPV |
|---|---|---|---|---|---|---|---|---|---|
Prediction in first year of life (0-12 mo) | |||||||||
Wolff | 2015 | Brain | Diffusion tensor imaging | Infant siblings (n = 270) and controls (n = 108) | 6 mo | Â | Â | 65-68% | 85-85% |
Hazlett | 2017 | Nature | Structural MRI | Infant siblings (n = 106) and controls (n=42) | 6-12 mo | 88% | 95% | 81% | 97% |
Emerson | 2017 | Science Translational Medicine | Functional MRI | Infant siblings (n = 59) | 6 mo | 81.80% | 100% | 100% | 96% |
Shen | 2013; 2017 | Brain; Biological Psychiatry | MRI (extra-axial cerebrospinal fluid) | Infant siblings (n = 33; n = 221) and controls (n = 22; n = 122) | 6 mo | 66-80% | 67-68% | 36-62% | 83-88% |
Bosl | 2018 | Scientific Reports | EEG entropy | Infant siblings (n = 99) and controls (n = 89) | 3 mo | 82% | 99% | 97% | - |
Gabard-Durnam | 2019 | Nature Communications | Frontal EEG spectral power | Infant siblings (n = 102) and controls (n = 69) | 3-12 mo | 82% | 86% | 72% | 92% |
Prediction in older infants/toddlers ( > 12 mo) | |||||||||
Perochon | 2023 | Nature Medicine | Digital behavioral phenotyping | General population (pediatric primary care; n = 475) | 17-36 mo | 87.80% | 80.80% | 97.80% | 40.60% |
Jones | 2023 | JAMA | Eye tracking | General population (autism specialty clinics; n = 475) | 16-30 | 71% | 80.70% | 76.20% | 76.20% |
Jones | 2023 | JAMA Network Open | Eye tracking in two cohorts (discovery and replication) | General population (autism specialty clinics; n = 1089 total, 719 in discovery, 370 in replication) | 16-45 | 80.6- | 82.3- | 81.9-87.3% | 81-85.4% |
81.9% | 89.9% | ||||||||
Active NIH grants replicating/extending prior work for prediction in infancy | |||||||||
Wetherby | 2018 | R01MH121364 | Mobile screening app | General population (pediatric primary care) | 2-24 mo | Â | |||
Pruett | 2019 | R01MH118362 | Structural and functional MRI | Infant siblings | 6 mo | ||||
Dawson | 2019 | R01MH121329 | Digital phenotyping via app | General population | 6-12 mo | ||||
Jones | 2019 | R01MH121363 | Eye tracking | General population (pediatric primary care) | 9 mo | ||||
Ozonoff | 2019 | R01MH121344 | Video coding | General population | 0-12 mo | ||||
Jeste | 2021 | R01MH121462 | EEG and eye tracking | Infant siblings | 6, 12 mo | ||||
Nelson | 2021 | R01NS120986 | Resting EEG | General population (pediatric primary care) | 4, 9, 12 mo | ||||
Sheinkopf | 2022 | R01MH121345 | Infant cry acoustics | General population | 0-12 mo | ||||