Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0630
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-08-03T06:08:21.032417Z
status: published
lens: trace
sector: health
headline: From machine learning to deep learning in attention deficit hyperactivity disorder diagnosis: A bibliometric analysis of global trends (2011-2024)
dek: Background The diagnosis of attention deficit hyperactivity disorder (ADHD) has traditionally relied on subjective clinical interviews. Recent years have witnessed a paradigm shift toward objective, data-driven diagnostics powered by artificial intelligence (AI). Objective This study provides a comprehensive bibliometric review of AI and machine learning (ML) applications in ADHD prediction to map the field's evolution, current trends, and future directions. Methods A structured search of the Scopus database ret…
gain_title: AI research is moving ADHD diagnosis away from subjective interviews toward objective, data-driven tools, with EEG emerging as preferred modality for recent models.
problem_title: High algorithmic accuracy in AI models for ADHD has not yet translated into routine clinical utility without advances in explainability and multimodal fusion.
trace_subject: AI-based objective diagnosis of ADHD
gain_reading: AI research is moving ADHD diagnosis away from subjective interviews toward objective, data-driven tools, with EEG emerging as preferred modality for recent models.
gain_evidence: paradigm shift toward objective, data-driven diagnostics powered by artificial intelligence (AI) | electroencephalography (EEG) has emerged as the preferred neuroimaging modality over functional magnetic resonance imaging (fMRI) in recent AI studies, driven by its cost-effectiveness and high temporal resolution
problem_reading: High algorithmic accuracy in AI models for ADHD has not yet translated into routine clinical utility without advances in explainability and multimodal fusion.
problem_evidence: Future research must prioritize explainable AI (XAI) and multimodal data fusion to translate high algorithmic accuracy into clinical utility
quick_read: A bibliometric review of 722 Scopus-indexed papers from 2011 to 2024 tracked how artificial intelligence has been applied to ADHD prediction. Using Python and VOSviewer, the authors found exponential growth peaking in 2023, a concentration of output in the United States and China, and a technological shift from support vector machines to deep learning with EEG becoming the favored data modality.

The shift matters because ADHD diagnosis has long depended on subjective clinical interviews, and objective AI tools could change clinical workflows if validated. Uncertainty remains about clinical adoption, as the authors conclude that explainable AI and multimodal data fusion are still needed to convert reported accuracy into usable psychiatric practice.
limitation: 
tag: Automated dual reading
key_points: Analyzed 722 Scopus publications from 2011 to 2024 using Python and VOSviewer. | Publication output peaked in 2023, indicating exponential growth phase. | Geographical distribution shows 130 papers from United States and 129 from China. | Technological trend shows transition from support vector machines to deep learning architectures.
rundown: A structured Scopus search retrieved 722 publications and used bibliometric indicators to assess annual production, geography, and technology. The analysis documents a research duopoly between the US and China and a peak in output in 2023.

The authors report a distinct move from traditional machine learning such as support vector machines to deep learning architectures, and a preference shift toward EEG over fMRI for cost and temporal resolution reasons.
sources:
- peer_reviewed | Applied Neuropsychology: Adult | https://doi.org/10.1080/23279095.2026.2707490 | 2026-08-01
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

81a27964afa69e2665dbf70fb364c0fa5713bde9ddb6d5a4121ab12ca07bfcdb

previous

0000000000000000000000000000000000000000000000000000000000000000