Towards a Near Infrared Spectroscopy-Based Estimation of Operator Attentional State

Derosière, Gérard; Dalhoumi, Sami; Perrey, Stéphane; Dray, Gérard; Ward, Tomás · 2014 · OpenAlex-citations

DOI: 10.1371/journal.pone.0092045

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the feasibility of using Near Infrared Spectroscopy (NIRS) combined with machine learning to detect decrements in operator attention during sustained cognitive tasks. Attention lapses pose significant risks to public safety, contributing to workplace accidents and traffic incidents. While reaction time (RT) is a standard behavioral indicator of attention decline, there is no established "gold standard" technology for real-time cognitive state monitoring. The authors aimed to determine if NIRS-measured cortical hemodynamic responses could reliably distinguish between states of full attention and reduced attention capacity, specifically comparing the efficacy of different brain regions and hemoglobin variables. Seven male participants performed a 30-minute sustained attention reaction time task, modeled after the psychomotor vigilance test. NIRS signals were recorded from the prefrontal cortex (PFC) and the right parietal area. The first 10 minutes of the task were classified as "full attention," while the final 10 minutes were classified as "attention decrement." Signal preprocessing included low-pass filtering, artifact removal, and z-normalization. A linear Support Vector Machine (SVM) algorithm was employed for two-class classification using leave-one-out cross-validation. Features included oxyhemoglobin (O2Hb) and deoxyhemoglobin (HHb) concentrations derived from the NIRS channels. Behavioral results confirmed that attention decrement occurred, evidenced by a significant increase in reaction time in the last 10 minutes compared to the first 10 minutes (p < 0.05, Cohen’s d = 0.7). Classification accuracy ranged from 65% to 90%. The highest accuracy was achieved using O2Hb signals rather than HHb signals. Crucially, signals from the right parietal area yielded superior classification performance (up to 90%) compared to those from the prefrontal cortex (77–89%). Combining both brain regions did not significantly improve accuracy beyond using the right parietal area alone. The findings demonstrate that NIRS-based classification can effectively detect attentional states, supporting the development of portable cognitive tracking technologies. The study highlights the right parietal area as a more sensitive biomarker for attention decrement than the traditionally studied prefrontal cortex. This suggests that future neuroimaging-based monitoring systems should prioritize parietal activity to maximize detection accuracy. The results imply that such technologies could facilitate real-time alerts for operators in high-stakes environments, potentially mitigating risks associated with cognitive fatigue.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• gaze based attention detection

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data