Table 3.

The proposed approach in comparison with related work.

Ref.	Detected Cognitive State(s)	Sample Size	Used Model(s)	Performance
[5]	Attention	24	Support Vector Machines (SVM)	76.8%
[8]	Four Attention states	4	NN	56.5–79.75%
[9]	Discrimination between working memory and recognition	10	SVM	79%
[19]	Four cognitive states related to an ambulatory subject	A total of 6000 data points taken from 1 subject	Ensemble Classifier	80%
[21]	Cognitive states related to viewing 4 images	26	Convolutional NN	79.9%
Proposed Approach	Three different levels (low, average, high) for the working memory and focused attention	86	NN and SVC	84% for WM and 81% for FA