User activity recognition system to improve the performance of environmental control interfaces: a pilot study with patients.
Arturo Bertomeu-Motos, Santiago Ezquerro, Juan A Barios and 6 others
PMID 30646915WHAT IT FOUND
In eight patients with neurological or spinal cord conditions, a prediction-assisted environment control interface let them complete more daily tasks in ten minutes than manual scanning.
Users did not report lower workload.
Key findings
01The AIDE mode let participants complete more tasks in the same time and spend less mean time per task than the manual mode (p < 0.001).
02The intention-prediction model had a global accuracy of about 87% on the ADL menu labels.
03Participants did not report lower subjective workload in the AIDE mode than in the manual mode.
STILL TO COME
How it was doneWhat they foundWhat 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
This was a pilot study with eight participants, so it cannot establish how the interface performs in a larger group. The system was tested only in a simulated home environment, not in the user's real environment. The AIDE mode combined EEG/EOG control, eye tracking, environment inputs, and a prediction model, so the contribution of any single component cannot be separated. Some tasks were excluded when participants became blocked by frustration or fatigue, so the reported performance may not include all attempted tasks. The prediction model confused Drink and Eat menus in around 50% of trials. Participants did not report lower subjective workload in the prediction-assisted mode.
Declared interests
Funding came from H2020 LEIT Information and Communication Technologies and Ministerio de Economía, Industria y Competitividad, Gobierno de España. The supplied text does not report author conflicts of interest.
The easy way to misread this
Do not conclude that this interface is ready for home use or that it reduces user burden. It was tested in only eight patients in a simulated home, and workload ratings did not show a subjective improvement in the prediction-assisted mode.