QStates Software

Rapid & Efficient Machine-Learning

QStates: Cognitive State Classification Using Machine Learning

QStates is a rapid and efficient machine learning software tool developed by QUASAR that uses quantitative EEG and heart rate variability data to assess cognitive and physiological states. Cognitive state assessment can be done in real-time and off-line with graphical displays of results.

QStates offers its users the flexibility to create their own models, which can be trained to classify any cognitive state for which there is a qEEG signature (e.g. cognitive workload, fatigue, engagement, or emotional states). Wearable Sensing and QUASAR scientists have validated the performance of greater than 90% accuracy for the states mental workload, engagement, and fatigue.

Training models are straightforward and fast: to create a cognitive state model, a user needs to collect a minimum of one minute of EEG data for each of the high and low state conditions (e.g. high workload vs. low workload).

QStates can classify up to 3 models simultaneously. The software automatically monitors data quality and rejects bad epochs. The program offers two different machine learning algorithm outputs: linear interpolation or probability density function. Outputs of state are given as values from 0-100 every 2 seconds. QStates offers automated summary tables generation. Data are saved in comma-separated value (.csv) format. QStates interfaces seamlessly with DSI-Streamer and the various DSI systems.

This is an example of real-time cognitive state classification for the models of engagement (left gauge) and cognitive workload (right 2 gauges):

EEG-Based Cognitive State Classification Using QStates

EEG spectral changes are being used for accurate estimation of alertness and cognitive workload (Makeig and Jung, 1995; Pope et al., 1995) and cognitive fatigue (Trejo, 2004). Furthermore, a number of studies have reported that theta is related to increases in attention, workload, memory load, and working memory performance, and that a large increase in alpha EEG activity precedes dozing off during a simple visual task (Torsvall and Akerstedt, 1988). EEG data have also been used to monitor the progress of trainees through skill levels or identify indices of skill acquisition. One group reported an increase in event-related alpha power that correlated with the amount of practice at a shooting task and suggested that it reflected a decrease in cortical activity associated with the reduced effort required with expertise (Kerick et al., 2004). Another group observed lower coherence associated with less cortico-cortical communication in expert marksmen compared to skilled shooters and attributed this difference to the decreased involvement of cognition with expertise (Deeny et al., 2003).
This is an example of real-time cognitive state classification for the models of engagement (left gauge) and cognitive workload (right 2 gauges):

QStates’ Cognitive Workload Performance

QStates

QStates is a software package developed by QUASAR that uses quantitative EEG and heart rate variability data for assessment of cognitive and physiological states. This machine learning algorithm first calculates several thousand spectral EEG features, then a partial least squares algorithm uses the most salient of these features as inputs to set weights of cognitive models based on the data collected during calibration runs of defined tasks (e.g. easy vs. hard tasks). Training a model requires as little as one minute of EEG data for each state (e.g. easy vs. hard), and is computationally expedient. The EEG data collected during experiments are then classified with these trained models to produce real-time cognitive state measures whose output ranges from 1 to 100, representing the probability that the EEG data fall into one state or the other (e.g. easy or hard). QUASAR’s fast-training algorithms allow for expedient calibration within minutes. These models typically produce average classification accuracies of greater than 90%. Furthermore, the models’ outputs track task difficulty reliably, correctly interpolating cognitive workload for tasks of intermediate difficulty compared to those used for training. Average classification accuracy of models for mental workload, engagement, and fatigue across 18 subjects.

Average mental workload model output on 18 subjects across varying task difficulty (average ± STD)