Quantifying the time for accurate EEG decoding of single value-based decisions

Tzovara, Athina; Chavarriaga, Ricardo; Lucia, Marzia De

doi:10.1016/j.jneumeth.2014.09.029

Cited by 9 publications

(8 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We (Goto et al, 2017) found that viewing a group of highly preferred (HP) consumer goods led to more positive N200, LPP and PSW amplitudes compared to viewing less preferred (LP) goods. These results are consistent with results showing that preferences for pictorial figures as well as food items could modulate electrophysiological activity as early as 150 ms after stimulus onset, and up to 800 ms (Harris et al, 2011; Tzovara et al, 2015). Further, Pozharliev et al (2015) did report effects of preference for luxury goods on the LPP, and Ma et al (2018) reported larger LPP amplitudes for prices that could lead to buying intentions.…”

Section: Introductionsupporting

confidence: 92%

Can Brain Waves Really Tell If a Product Will Be Purchased? Inferring Consumer Preferences From Single-Item Brain Potentials

Goto

Lim

Shee

et al. 2019

Front. Integr. Neurosci.

View full text Add to dashboard Cite

Recent research has shown that event-related brain potentials (ERPs) recorded while participants view lists of different consumer goods can be modulated by their preferences toward these products. However, it remains largely unknown whether ERP activity specific to a single consumer item can be informative about whether or not this item will be preferred in a shopping context. In this study, we examined whether single-item ERPs could reliably predict consumer preferences toward specific consumer goods. We recorded scalp EEG from 40 participants while they were viewing pictures of consumer goods and we subsequently asked them to indicate their preferences for each of these items. Replicating previous results, we found that ERP activity averaged over the six most preferred products was significantly differentiated from ERP activity averaged across the six least preferred products for three ERP components: The N200, the late positive potential (LPP) and positive slow waves (PSW). We also found that using single-item ERPs to infer behavioral preferences about specific consumer goods led to an overall predictive accuracy of 71%, although this figure varied according to which ERPs were targeted. Later positivities such as the LPP and PSW yielded relatively higher predictive accuracy rates than the frontal N200. Our results suggest that ERPs related to single consumer items can be relatively accurate predictors of behavioral preferences depending on which type of ERP effects are chosen by the researcher, and ultimately on the level of prediction errors that users choose to tolerate.

show abstract

Section: Introductionsupporting

confidence: 92%

Can Brain Waves Really Tell If a Product Will Be Purchased? Inferring Consumer Preferences From Single-Item Brain Potentials

Goto

Lim

Shee

et al. 2019

Front. Integr. Neurosci.

View full text Add to dashboard Cite

show abstract

“…One cluster of activity occurred frontocentrally directly after the modelfree prediction error. Studies by Tzovara, et al (2015), Fischer and Ullsperger (2013) and Hunt, et al (2012) (using MEG), employing tasks in which variably valuable choices were presented at the start of a trial have also successfully demonstrated neural signatures of Q values. Also, in an experiment employing the same two step task as that used here, Eppinger, et al (2017) showed state action Q values at state 2 at a comparable latency (400 ms) to ourselves, though more parietally.…”

Section: The Demonstration Of Model-based Rpes Does Not Resolve the Qmentioning

confidence: 96%

Model-free and model-based reward prediction errors in EEG

et al. 2018

View full text Add to dashboard Cite

Learning theorists posit two reinforcement learning systems: model-free and model-based. Model-based learning incorporates knowledge about structure and contingencies in the world to assign candidate actions with an expected value. Model-free learning is ignorant of the world's structure; instead, actions hold a value based on prior reinforcement, with this value updated by expectancy violation in the form of a reward prediction error. Because they use such different learning mechanisms, it has been previously assumed that model-based and model-free learning are computationally dissociated in the brain. However, recent fMRI evidence suggests that the brain may compute reward prediction errors to both model-free and model-based estimates of value, signalling the possibility that these systems interact. Because of its poor temporal resolution, fMRI risks confounding reward prediction errors with other feedback-related neural activity. In the present study, EEG was used to show the presence of both model-based and model-free reward prediction errors and their place in a temporal sequence of events including state prediction errors and action value updates. This demonstration of model-based prediction errors questions a long-held assumption that model-free and model-based learning are dissociated in the brain.

show abstract

“…Subjective or cognitive performance measures have been employed in the analysis of single-trial evoked potential data to separate noise variance from the meaningful information about the perceptual decisions occurring with each stimulus. Mapping behavioural data onto brain electrical activity requires a statistical model, such as a multiple linear regression model (Ratcliff et al 2009 ; Rousselet et al 2011 ), support vector machines (Schulz et al 2012 ), multivariate decoding based on accumulation of topographic activity (Tzovara et al 2015 ), or a linear fixed effects model (Michail et al 2016 ). While single-trial analysis of evoked potentials was previously performed using a variety of methods, the multiple regression analysis utilised here fitted the objective to associate multiple aspects of subjective pain experience with the P2 component whilst accounting for possible inter-correlations between predictors.…”

Section: Methodsmentioning

confidence: 99%

Neural Mechanisms of Attentional Switching Between Pain and a Visual Illusion Task: A Laser Evoked Potential Study

et al. 2017

View full text Add to dashboard Cite

Previous studies demonstrated that pain induced by a noxious stimulus during a distraction task is affected by both stimulus-driven and goal-directed processes which interact and change over time. The purpose of this exploratory study was to analyse associations of aspects of subjective pain experience and engagement with the distracting task with attention-sensitive components of noxious laser-evoked potentials (LEPs) on a single-trial basis. A laser heat stimulus was applied to the dorsum of the left hand while subjects either viewed the Rubin vase-face illusion (RVI), or focused on their pain and associated somatosensory sensations occurring on their stimulated hand. Pain-related sensations occurring with every laser stimulus were evaluated using a set of visual analogue scales. Factor analysis was used to identify the principal dimensions of pain experience. LEPs were correlated with subjective aspects of pain experience on a single-trial basis using a multiple linear regression model. A positive LEP component at the vertex electrodes in the interval 294–351 ms (P2) was smaller during focusing on RVI than during focusing on the stimulated hand. Single-trial amplitude variations of the P2 component correlated with changes in Factor 1, representing essential aspects of pain, and inversely with both Factor 2, accounting for anticipated pain, and the number of RVI figure reversals. A source dipole located in the posterior region of the cingulate cortex was the strongest contributor to the attention-related single-trial variations of the P2 component. Instantaneous amplitude variations of the P2 LEP component during switching attention towards pain in the presence of a distracting task are related to the strength of pain experience, engagement with the task, and the level of anticipated pain. Results provide neurophysiological underpinning for the use of distraction analgesia acute pain relief.

show abstract

Quantifying the time for accurate EEG decoding of single value-based decisions

Cited by 9 publications

References 54 publications

Can Brain Waves Really Tell If a Product Will Be Purchased? Inferring Consumer Preferences From Single-Item Brain Potentials

Can Brain Waves Really Tell If a Product Will Be Purchased? Inferring Consumer Preferences From Single-Item Brain Potentials

Model-free and model-based reward prediction errors in EEG

Neural Mechanisms of Attentional Switching Between Pain and a Visual Illusion Task: A Laser Evoked Potential Study

Contact Info

Product

Resources

About