Single-trial stimuli classification from detected P300 for augmented Brain–Computer Interface: A deep learning approach

Leoni, Jessica; Strada, Silvia; Tanelli, Mara; Brusa, Alessandra; Proverbio, Alice Mado

doi:10.1016/j.mlwa.2022.100393

Cited by 15 publications

(10 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This was shown in an interesting investigation, where a set of six pictures of familiar objects was correctly classified with an AI classificator based on the P300 response in a small group of healthy subjects and three post-stroke people with an accuracy of 91.79% and 89.68% for healthy and disabled subjects, respectively (Cortez et al, 2021). Higher levels of accuracy were obtained in healthy controls by Proverbio and coworkers by automatically classifying electrical signals relative to 12 perceptual categories by means of statistical analyses , and supervised machine-learning systems (Leoni et al, 2021(Leoni et al, , 2022 applied to ERPs (Leoni et al, 2021) and EEG single-trials (Leoni et al, 2022). Apart from this literature, we are not aware of BCI studies currently using pictorial communication systems to investigate sensations, volitional and affective states.…”

Section: Discussionmentioning

confidence: 99%

Validation of a pictionary-based communication tool for assessing physiological needs and motivational states: the PAIN set

Proverbio

Pischedda

2023

Front. Cognit.

Self Cite

View full text Add to dashboard Cite

This Pictionary is suitable for communicating with individual unable to interact (locked-in syndrome, LIS) through the development of Brain Computer Interface (BCI) systems. It includes 60 validated, easy-to-understand illustrated plates depicting adults in various situations affecting their physiological or psychological state. The drawings are in color and represent persons of both sexes and various ethnicities. Twenty participants were interviewed to appropriately design the Pictionary. An additional group of 50 healthy adults (25 women and 25 men) aging 18–33 years was recruited to validate the pictogram corpus. Their schooling levels ranged from middle school to master's degrees. Participants were presented with five runs of randomly mixed pictograms illustrating 12 different motivational states, including primary and secondary needs, affective states, and somatosensory sensations (with five variants for each category). They had to precisely identify the motivational category illustrated (e.g., “Feeling pain” or “Being hungry”) while also providing information about its clarity and unambiguity on a Likert scale. Statistical analyses provided evidence of the strong communicative effectiveness of the illustrations (rated on average 2.7, on a 0 to 3 scale), with an accuracy of 98.4%. The PAIN set could be a valuable communication tool for individuals with LIS, as well as any clinical population lacking verbal communication skills. Its main purpose is to generate electrophysiological markers of internal mental states to be automatically classified by BCI systems.

show abstract

Section: Discussionmentioning

confidence: 99%

Validation of a pictionary-based communication tool for assessing physiological needs and motivational states: the PAIN set

Proverbio

Pischedda

2023

Front. Cognit.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Clear evidences have been provided of reliable electrophysiological markers for motor imagery (e.g., Milanés-Hermosilla et al, 2021 ; Mattioli et al, 2022 ), and, to a lesser extent, perceptual and cognitive imagery ( Cai et al, 2013 ; Leoni et al, 2021 , 2022 ; Proverbio et al, 2022 ), for communication with brain computer interface (BCI) systems ( Ash and Benson, 2018 ). In contrast, not much is known about the electrophysiological markers of motivational imagery (craves, wills, needs and desires), despite the fact that this particular type of mental content is valuable for interacting with patients with disorders of consciousness, such as coma or locked-in syndrome.…”

Section: Introductionmentioning

confidence: 99%

Measuring brain potentials of imagination linked to physiological needs and motivational states

Proverbio

Pischedda

2023

Front. Hum. Neurosci.

View full text Add to dashboard Cite

IntroductionWhile EEG signals reflecting motor and perceptual imagery are effectively used in brain computer interface (BCI) contexts, little is known about possible indices of motivational states. In the present study, electrophysiological markers of imagined motivational states, such as craves and desires were investigated.MethodsEvent-related potentials (ERPs) were recorded in 31 participants during perception and imagery elicited by the presentation of 360 pictograms. Twelve micro-categories of needs, subdivided into four macro-categories, were considered as most relevant for a possible BCI usage, namely: primary visceral needs (e.g., hunger, linked to desire of food); somatosensory thermal and pain sensations (e.g., cold, linked to desire of warm), affective states (e.g., fear: linked to desire of reassurance) and secondary needs (e.g., desire to exercise or listen to music). Anterior N400 and centroparietal late positive potential (LPP) were measured and statistically analyzed.ResultsN400 and LPP were differentially sensitive to the various volition stats, depending on their sensory, emotional and motivational poignancy. N400 was larger to imagined positive appetitive states (e.g., play, cheerfulness) than negative ones (sadness or fear). In addition, N400 was of greater amplitude during imagery of thermal and nociceptive sensations than other motivational or visceral states. Source reconstruction of electromagnetic dipoles showed the activation of sensorimotor areas and cerebellum for movement imagery, and of auditory and superior frontal areas for music imagery.DiscussionOverall, ERPs were smaller and more anteriorly distributed during imagery than perception, but showed some similarity in terms of lateralization, distribution, and category response, thus indicating some overlap in neural processing, as also demonstrated by correlation analyses. In general, anterior frontal N400 provided clear markers of subjects’ physiological needs and motivational states, especially cold, pain, and fear (but also sadness, the urgency to move, etc.), than can signal life-threatening conditions. It is concluded that ERP markers might potentially allow the reconstruction of mental representations related to various motivational states through BCI systems.

show abstract

“…Furthermore, it would be impractical to collect an enormously large number of single trials to generate large ensembles of high SNR ERPs to form training and test sets. The most obvious solution is to use single trails directly, without averaging, as attempted in the design of customized brain computer interfaces (BCIs) which are typically controlled by the presentation of a single stimulus, that is, by single trials [28][29][30][31][32][33]. In general, irrespective of the application, high classification accuracies cannot be expected with single trials due to the poor SNR.…”

Section: Introductionmentioning

confidence: 99%

Analyses of Subsample Averaging of Single Trials for Event-Related Potential Classification

Chen¹,

Gupta²

2023

Preprint

View full text Add to dashboard Cite

Event-related potentials (ERPs) are estimated by averaging time-locked single trial electroencephalography (EEG) signals in response to specific events or stimuli. Classifying ERPs accurately is a challenge because (a) single trials have poor signal-to-noise-ratios (SNRs) and (b) it is difficult to collect large single trial ensembles to generate high SNR ERPs for classifier training and testing. The m-subsample averaging (m-SA) strategy which generates small-sample ERPs by repeated averaging of a small number of single trials drawn without replacement, has been proposed as a solution to the two problems. An ERP formed by averaging m single trials is referred to as an m-ERP where m is referred to as the averaging parameter. In this study, we conduct thorough analyses of m-SA and focus on issues not addressed in previous studies to better understand the beneficial properties of m-SA and to further support its application for ERP classification. Specifically, we (a) analyze the improvement in SNR as a function of m using the mean-root-mean-square SNR and visual analyses of m-ERP plots with confidence intervals, (b) analyze the improvement in interclass separation as a function of m, (c) determine how the SNR and interclass separation analyses can help to select the averaging parameter m, (d) determine the number of distinct m-ERPs that can be drawn from a single-trial ensemble, and (e) determine several probabilities related to the generation of distinct m-ERPs. Furthermore, an extensive set of experiments are designed to analyze the performance of support vector machine and convolution neural network classifiers employing m-SA with various combinations of the averaging parameters used for generating the training and test sets. The results confirm that ERPs can be classified accurately using small subsample averaging. Most importantly, it is concluded that m-SA can be deployed in practice to accurately classify ERPs in brain activity research and in clinical applications without having to collect a prohibitively large number of single trials.

show abstract

Single-trial stimuli classification from detected P300 for augmented Brain–Computer Interface: A deep learning approach

Cited by 15 publications

References 37 publications

Validation of a pictionary-based communication tool for assessing physiological needs and motivational states: the PAIN set

Validation of a pictionary-based communication tool for assessing physiological needs and motivational states: the PAIN set

Measuring brain potentials of imagination linked to physiological needs and motivational states

Analyses of Subsample Averaging of Single Trials for Event-Related Potential Classification

Contact Info

Product

Resources

About