Investigating sensory response to physical discomfort in children with autism spectrum disorder using near-infrared spectroscopy.

Self-reporting of pain can be difficult in populations with communication challenges or atypical sensory processing, such as children with autism spectrum disorder (ASD). Consequently, pain can go untreated. An objective method to identify discomfort would be valuable to individuals unable to express or recognize their own bodily distress. Near-infrared spectroscopy (NIRS) is a brain-imaging modality that is suited for this application. We evaluated the potential of detecting a cortical response to discomfort in the ASD population using NIRS. Using a continuous-wave spectrometer, prefrontal and parietal measures were collected from 15 males with ASD and 7 typically developing (TD) males 10-15 years of age. Participants were exposed to a noxious cold stimulus by immersing their hands in cold water and tepid water as a baseline task. Across all participants, the magnitude and timing of the cold and tepid water-induced brain responses were significantly different (p < 0.001). The effect of the task on the brain response depended on the study group (group x task: p < 0.001), with the ASD group exhibiting a blunted response to the cold stimulus. Findings suggest that NIRS may serve as a tool for objective pain assessment and atypical sensory processing.

Development of a Ternary Near-Infrared Spectroscopy Brain-Computer Interface: Online Classification of Verbal Fluency Task, Stroop Task and Rest.

The majority of proposed NIRS-BCIs has considered binary classification. Studies considering high-order classification problems have yielded average accuracies that are less than favorable for practical communication. Consequently, there is a paucity of evidence supporting online classification of more than two mental states using NIRS. We developed an online ternary NIRS-BCI that supports the verbal fluency task (VFT), Stroop task and rest. The system utilized two sessions dedicated solely to classifier training. Additionally, samples were collected prior to each period of online classification to update the classifier. Using a continuous-wave spectrometer, measurements were collected from the prefrontal and parietal cortices while 11 able-bodied adult participants were cued to perform one of the two cognitive tasks or rests. Each task was used to indicate the desire to select a particular letter on a scanning interface, while rest avoided selection. Classification was performed using 25 iteration of bagging with a linear discriminant base classifier. Classifiers were trained on 10-dimensional feature sets. The BCI's classification decision was provided as feedback. An average online classification accuracy of [Formula: see text]% was achieved, representing an ITR of [Formula: see text] bits/min. The results demonstrate that online communication can be achieved with a ternary NIRS-BCI that supports VFT, Stroop task and rest. Our findings encourage continued efforts to enhance the ITR of NIRS-BCIs.

Development and testing an online near-infrared spectroscopy brain-computer interface tailored to an individual with severe congenital motor impairments.

PURPOSE: For non-verbal individuals, brain-computer interfaces (BCIs) are a potential means of communication. Near-infrared spectroscopy (NIRS) is a brain-monitoring modality that has been considered for BCIs. To date, limited NIRS-BCI testing has involved online classification, particularly with individuals with severe motor impairments. MATERIALS AND METHODS: We tested an online NIRS-BCI developed for a non-verbal individual with severe congenital motor impairments. The binary BCI differentiated categorical verbal fluency task (VFT) performance and rest using prefrontal measurements. The participant attended five sessions, the last two of which were online with classification feedback. RESULTS: An online classification accuracy of 63.33% was achieved using a linear discriminant classifier trained on a four-dimensional feature set. An offline, cross-validation analysis of all data yielded an optimal adjusted classification accuracy of 66.6 ± 9.11%. Inconsistent functional responses, contradictory effects of feedback, participant fatigue and motion artefacts were identified as challenges to online classification specific to this participant. CONCLUSIONS: Results suggest potential in using an NIRS-BCI controlled by the VFT in instances of severe congenital impairments. Further testing with users with severe disabilities is necessary. Implications for Rehabilitation Brain-computer interfaces (BCIs) can provide a non-motor based means of communication for individuals with severe motor impairments. Near-infrared spectroscopy (NIRS) is a haemodynamic-based brain-imaging modality used in BCIs. To date, NIRS-BCIs have not been thoroughly tested with potential target users. This case study shows that NIRS-BCIs may offer a means of practical communication for individuals with severe congenital impairments and continued exploration is advisable.

Single-trial classification of near-infrared spectroscopy signals arising from multiple cortical regions.

Near-infrared spectroscopy (NIRS) brain-computer interface (BCI) studies have primarily made use of measurements taken from a single cortical area. In particular, the anterior prefrontal cortex has been the key area used for detecting higher-level cognitive task performance. However, mental task execution typically requires coordination between several, spatially-distributed brain regions. We investigated the value of expanding the area of interrogation to include NIRS measurements from both the prefrontal and parietal cortices to decode mental states. Hemodynamic activity was monitored at 46 locations over the prefrontal and parietal cortices using a continuous-wave near-infrared spectrometer while 11 able-bodied adults rested or performed either the verbal fluency task (VFT) or Stroop task. Offline classification was performed for the three possible binary problems using 25 iterations of bagging with a linear discriminant base classifier. Classifiers were trained on a 10 dimensional feature set. When all 46 measurement locations were considered for classification, average accuracies of 80.4±7.0%, 82.4±7.6%, and 82.8±5.9% in differentiating VFT vs rest, Stroop vs rest and VFT vs Stroop, respectively, were obtained. Relative to using measurements from the anterior PFC alone, an overall average improvement of 11.3% was achieved. Utilizing NIRS measurements from the prefrontal and parietal cortices can be of value in classifying mental states involving working memory and attention. NIRS-BCI accuracies may be improved by incorporating measurements from several, distinct cortical regions, rather than a single area alone. Further development of an NIRS-BCI supporting combinations of VFT, Stroop task and rest states is also warranted.

Dynamic topographical pattern classification of multichannel prefrontal NIRS signals: II. Online differentiation of mental arithmetic and rest.

OBJECTIVE: Near-infrared spectroscopy (NIRS) has recently gained attention as a modality for brain-computer interfaces (BCIs), which may serve as an alternative access pathway for individuals with severe motor impairments. For NIRS-BCIs to be used as a real communication pathway, reliable online operation must be achieved. Yet, only a limited number of studies have been conducted online to date. These few studies were carried out under a synchronous paradigm and did not accommodate an unconstrained resting state, precluding their practical clinical implication. Furthermore, the potentially discriminative power of spatiotemporal characteristics of activation has yet to be considered in an online NIRS system. APPROACH: In this study, we developed and evaluated an online system-paced NIRS-BCI which was driven by a mental arithmetic activation task and accommodated an unconstrained rest state. With a dual-wavelength, frequency domain near-infrared spectrometer, measurements were acquired over nine sites of the prefrontal cortex, while ten able-bodied participants selected letters from an on-screen scanning keyboard via intentionally controlled brain activity (using mental arithmetic). Participants were provided dynamic NIR topograms as continuous visual feedback of their brain activity as well as binary feedback of the BCI's decision (i.e. if the letter was selected or not). To classify the hemodynamic activity, temporal features extracted from the NIRS signals and spatiotemporal features extracted from the dynamic NIR topograms were used in a majority vote combination of multiple linear classifiers. MAIN RESULTS: An overall online classification accuracy of 77.4 ± 10.5% was achieved across all participants. The binary feedback was found to be very useful during BCI use, while not all participants found value in the continuous feedback provided. SIGNIFICANCE: These results demonstrate that mental arithmetic is a potent mental task for driving an online system-paced NIRS-BCI. BCI feedback that reflects the classifier's decision has the potential to improve user performance. The proposed system can provide a framework for future online NIRS-BCI development and testing.

A review of past and future near-infrared spectroscopy brain computer interface research at the PRISM lab.

Single-trial classification of near-infrared spectroscopy (NIRS) signals for brain-computer interface (BCI) applications has recently gained much attention. This paper reviews research in this area conducted at the PRISM lab (University of Toronto) to date, as well as directions for future work. Thus far, research has included classification of hemodynamic changes induced by the performance of various mental tasks in both offline and online settings, as well as offline classification of cortical changes evoked by different affective states. The majority of NIRS-BCI work has only involved able-bodied individuals. However, preliminary work involving individuals from target BCI-user populations is also underway. In addition to further testing with users with severe disabilities, ongoing and future research will focus on enhancing classification accuracies, communication speed and user experience.

Dynamic topographical pattern classification of multichannel prefrontal NIRS signals.

OBJECTIVE: Near-infrared spectroscopy (NIRS) is an optical imaging technique that has recently been considered for brain-computer interface (BCI) applications. To date, NIRS-BCI studies have primarily made use of temporal features of brain activity, derived from the time-course of optical signals measured from discrete locations, to differentiate mental states. However, functional brain imaging studies have indicated that the spatial distribution of haemodynamic activity is also rich in information. Thus, the progression of a response over both time and space may be valuable to brain state classification. In this paper, we investigate the implication of including spatiotemporal features in the single-trial classification of haemodynamic events for a two-class problem by exploiting this information from dynamic NIR topograms. APPROACH: The value of spatiotemporal information was explored through a comparative analysis of four different classification schemes performed on multichannel NIRS data collected from the prefrontal cortex during a mental arithmetic activation task and rest. Employing a linear discriminant classifier, data were analysed using spatiotemporal features, temporal features, and a collective pool of spatiotemporal and temporal features. We also considered a majority vote combination of three classifiers; each established using one of the above feature sets. Lastly, two separate task durations (20 and 10 s) were considered for feature extraction. MAIN RESULTS: With features from the longer task interval, the highest overall classification accuracy was achieved using the majority voting classifier (76.1 ± 8.4%), which was greater than the accuracy obtained using temporal features alone (73.5 ± 8.5%) (F3,144 = 7.04, p = 0.0002). While results from the shorter task duration were lower overall, the classifier employing only spatiotemporal features (with an average accuracy of 67.9 ± 9.3%) achieved a higher average accuracy than the rate obtained using only temporal features (64.4 ± 8.4%) (F3,144 = 18.58, p < 10(-4)). SIGNIFICANCE: Collectively, these results suggest that spatiotemporal information can be of value in the analysis of functional NIRS data, and improved classification rates may be obtained in future NIRS-BCI applications with the inclusion of this information.