Augmented Reality and Artificial Intelligence for the Assessment and Rehabilitation of Spatial Neglect: A Systematic Review.
Shaojun Li, Benjamin Chong, Ghazal Mehri-Kakavand and 6 others
PMID 42080328WHAT IT FOUND
Augmented reality and AI show promise for assessing spatial neglect, with some diagnostic accuracy exceeding standard tests.
However, evidence for rehabilitation effectiveness is very low, with only one small trial. Current tools lack personalization and real-world validation.
Key findings
01AI assessment methods demonstrated high diagnostic accuracy (AUC > 0.80 to 0.95) in detecting spatial neglect, often matching or exceeding paper-and-pencil tests.
02Evidence for AR rehabilitation effectiveness is very low, with only one small RCT showing superiority over standard therapy on 3 of 5 measures.
03Most included studies had low ecological validity and minimal personalization, relying on rule-based adjustments rather than adaptive AI modeling.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for OTsWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The review included only 15 studies, with a very small total sample size for rehabilitation (n=20 in the only RCT). All included studies were from high-income countries, limiting generalizability to other healthcare settings. Most AI models were validated only via internal cross-validation without external testing, raising concerns about overfitting and real-world applicability. There was substantial heterogeneity in how spatial neglect was diagnosed and measured across studies, making direct comparisons difficult. The search excluded gray literature and studies published after August 2025, potentially missing recent developments.
Declared interests
The authors declared no potential conflicts of interest. The work was partially supported by grants from the Health Research Council of New Zealand, the Marsden Fund, and the Royal Society of New Zealand.
The easy way to misread this
Do not interpret the high diagnostic accuracy of AI tools as evidence they are ready for clinical use. Most models were tested only on small, single-center datasets without external validation, meaning they may not perform reliably in your practice or with different patient populations.
Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →