Abstract
Multivariate pattern classification methods are increasingly applied to neuroimaging data in the context of both fundamental research and in brain-computer interfacing approaches. Such methods provide a framework for interpreting measurements made at the single-trial level with respect to a set of two or more distinct mental states. Here, we define an approach in which the output of a binary classifier trained on data from an auditory mismatch paradigm can be used for online tracking of perception and as a neurofeedback signal. The auditory mismatch paradigm is known to induce distinct perceptual states related to the presentation of high- and low-probability stimuli, which are reflected in event-related potential (ERP) components such as the mismatch negativity (MMN). The first part of this paper illustrates how pattern classification methods can be applied to data collected in an MMN paradigm, including discussion of the optimization of preprocessing steps, the interpretation of features and how the performance of these methods generalizes across individual participants and measurement sessions. We then go on to show that the output of these decoding methods can be used in online settings as a continuous index of single-trial brain activation underlying perceptual discrimination. We conclude by discussing several potential domains of application, including neurofeedback, cognitive monitoring and passive brain-computer interfaces.
Highlights
The ability to non-invasively measure real-time changes in the patterns of brain activity underlying important perceptual and cognitive processes has led to breakthroughs in areas that were until recently the domain of science fiction
In contrast to the averaging methods often used to investigate brain responses measured in EEG, fMRI and other neuroimaging modalities, real-time tracking methods enable researchers to monitor the ongoing dynamics of brain activity as individuals perform different cognitive or behavioral tasks, to use brain responses as a control signal in a brain-computer interface (BCI)
While we would obviously expect the mismatch negativity (MMN) component to contribute to classifier performance in these analyses, it is not the only component of the auditory event-related potential (ERP) modulated during the presentation of a deviant trial
Summary
The ability to non-invasively measure real-time changes in the patterns of brain activity underlying important perceptual and cognitive processes has led to breakthroughs in areas that were until recently the domain of science fiction. Common to many of these approaches is the use of multivariate pattern classification methods, or so-called decoding approaches (Haynes and Rees, 2006; van Gerven et al, 2009; Blankertz et al, 2011) These machine learning techniques provide a means for making predictions about the mental state of a user on the basis of single-trial neuroimaging data. Given sufficient amounts of data, a classifier trained on such a dataset will learn to assign importance to specific features of the data corresponding to P300 responses elicited in individual trials while ignoring other features unrelated to the two classes of interest. In the context of auditory perception, similar sequences are used to elicit another ERP component: the MMN response These “oddball” sequences contain frequent standard trials and rare deviant trials, each corresponding to a different type of sound. It has been shown that decoding performance reflects differences in categorical speech perception by native and non-native speakers (Brandmeyer et al, 2013), and that decoding analyses can be used to predict survival rates in comatose patients (Tzovara et al, 2013)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
