Alberto Llera

On the use of interaction error potentials for adaptive brain computer interfaces

We propose an adaptive classification method for the Brain Computer Interfaces (BCI) which uses Interaction Error Potentials (IErrPs) as a reinforcement signal and adapts the classifier parameters when an error is detected. We analyze the quality of the proposed approach in relation to the misclassification of the IErrPs. In addition we compare static versus adaptive classification performance using artificial and MEG data. We show that the proposed adaptive framework significantly improves the static classification methods.

Adaptive classification on brain-computer interfaces using reinforcement signals

We introduce a probabilistic model that combines a classifier with an extra reinforcement signal (RS) encoding the probability of an erroneous feedback being delivered by the classifier. This representation computes the class probabilities given the task related features and the reinforcement signal. Using expectation maximization (EM) to estimate the parameter values under such a model shows that some existing adaptive classifiers are particular cases of such an EM algorithm. Further, we present a new algorithm for adaptive classification, which we call constrained means adaptive classifier, and show using EEG data and simulated RS that this classifier is able to significantly outperform state-of-the-art adaptive classifiers.

Adaptive multiclass classification for brain computer interfaces

We consider the problem of multiclass adaptive classification for brain-computer interfaces and propose the use of multiclass pooled mean linear discriminant analysis (MPMLDA), a multiclass generalization of the adaptation rule introduced by Vidaurre, Kawanabe, von Bünau, Blankertz, and Müller (2010) for the binary class setting. Using publicly available EEG data sets and tangent space mapping (Barachant, Bonnet, Congedo, & Jutten, 2012) as a feature extractor, we demonstrate that MPMLDA can significantly outperform state-of-the-art multiclass static and adaptive methods. Furthermore, efficient learning rates can be achieved using data from different subjects.

Integrated analysis of gray and white matter alterations in attention-deficit/hyperactivity disorder

Background Magnetic resonance imaging (MRI) is able to provide detailed insights into the structural organization of the brain, e.g., by means of mapping brain anatomy and white matter microstructure. Understanding interrelations between MRI modalities, rather than mapping modalities in isolation, will contribute to unraveling the complex neural mechanisms associated with neuropsychiatric disorders as deficits detected across modalities suggest common underlying mechanisms. Here, we conduct a multimodal analysis of structural MRI modalities in the context of attention-deficit/hyperactivity disorder (ADHD). Methods Gray matter volume, cortical thickness, surface areal expansion estimates, and white matter diffusion indices of 129 participants with ADHD and 204 participants without ADHD were entered into a linked independent component analysis. This data-driven analysis decomposes the data into multimodal independent components reflecting common inter-subject variation across imaging modalities. Results ADHD severity was related to two multimodal components. The first component revealed smaller prefrontal volumes in participants with more symptoms, co-occurring with abnormal white matter indices in prefrontal cortex. The second component demonstrated decreased orbitofrontal volume as well as abnormalities in insula, occipital, and somato-sensory areas in participants with more ADHD symptoms. Conclusions Our results replicate and extend previous unimodal structural MRI findings by demonstrating that prefrontal, parietal, and occipital areas, as well as fronto-striatal and fronto-limbic systems are implicated in ADHD. By including multiple modalities, sensitivity for between-participant effects is increased, as shared variance across modalities is modeled. The convergence of modality-specific findings in our results suggests that different aspects of brain structure share underlying pathophysiology and brings us closer to a biological characterization of ADHD.

Thresholding functional connectomes by means of mixture modeling

&NA; Functional connectivity has been shown to be a very promising tool for studying the large‐scale functional architecture of the human brain. In network research in fMRI, functional connectivity is considered as a set of pair‐wise interactions between the nodes of the network. These interactions are typically operationalized through the full or partial correlation between all pairs of regional time series. Estimating the structure of the latent underlying functional connectome from the set of pair‐wise partial correlations remains an open research problem though. Typically, this thresholding problem is approached by proportional thresholding, or by means of parametric or non‐parametric permutation testing across a cohort of subjects at each possible connection. As an alternative, we propose a data‐driven thresholding approach for network matrices on the basis of mixture modeling. This approach allows for creating subject‐specific sparse connectomes by modeling the full set of partial correlations as a mixture of low correlation values associated with weak or unreliable edges in the connectome and a sparse set of reliable connections. Consequently, we propose to use alternative thresholding strategy based on the model fit using pseudo‐False Discovery Rates derived on the basis of the empirical null estimated as part of the mixture distribution. We evaluate the method on synthetic benchmark fMRI datasets where the underlying network structure is known, and demonstrate that it gives improved performance with respect to the alternative methods for thresholding connectomes, given the canonical thresholding levels. We also demonstrate that mixture modeling gives highly reproducible results when applied to the functional connectomes of the visual system derived from the n‐back Working Memory task in the Human Connectome Project. The sparse connectomes obtained from mixture modeling are further discussed in the light of the previous knowledge of the functional architecture of the visual system in humans. We also demonstrate that with use of our method, we are able to extract similar information on the group level as can be achieved with permutation testing even though these two methods are not equivalent. We demonstrate that with both of these methods, we obtain functional decoupling between the two hemispheres in the higher order areas of the visual cortex during visual stimulation as compared to the resting state, which is in line with previous studies suggesting lateralization in the visual processing. However, as opposed to permutation testing, our approach does not require inference at the cohort level and can be used for creating sparse connectomes at the level of a single subject. HighlightsSparse functional connectomes are useful in analyzing and interpreting fMRI data.We propose thresholding by means of mixture modeling and control of FDR.We benchmark the approach on synthetic fMRI data against established methods.We apply the method to the resting state and working memory task datasets from HCP500.Results are reproducible on synthetic data and interpretable on experimental data.

Quantitative analysis of task selection for brain–computer interfaces

OBJECTIVE To assess quantitatively the impact of task selection in the performance of brain-computer interfaces (BCI). APPROACH We consider the task-pairs derived from multi-class BCI imagery movement tasks in three different datasets. We analyze for the first time the benefits of task selection on a large-scale basis (109 users) and evaluate the possibility of transferring task-pair information across days for a given subject. MAIN RESULTS Selecting the subject-dependent optimal task-pair among three different imagery movement tasks results in approximately 20% potential increase in the number of users that can be expected to control a binary BCI. The improvement is observed with respect to the best task-pair fixed across subjects. The best task-pair selected for each subject individually during a first day of recordings is generally a good task-pair in subsequent days. In general, task learning from the user side has a positive influence in the generalization of the optimal task-pair, but special attention should be given to inexperienced subjects. SIGNIFICANCE These results add significant evidence to existing literature that advocates task selection as a necessary step towards usable BCIs. This contribution motivates further research focused on deriving adaptive methods for task selection on larger sets of mental tasks in practical online scenarios.

Disentangling common from specific processing across tasks using task potency

&NA; When an individual engages in a task, the associated evoked activities build upon already ongoing activity, shaped by an underlying functional connectivity baseline (Fox et al., 2009; Smith et al., 2009; Tavor et al., 2016). Building on the idea that rest represents the brains full functional repertoire, we here incorporate the idea that task‐induced functional connectivity modulations ought to be task‐specific with respect to their underlying resting state functional connectivity. Various metrics such as clustering coefficient or average path length have been proposed to index processing efficiency, typically from single fMRI session data. We introduce a framework incorporating task potency, which provides direct access to task‐specificity by enabling direct comparison between task paradigms. In particular, to study functional connectivity modulations related to cognitive involvement in a task we define task potency as the amplitude of a connectivity modulation away from its baseline functional connectivity architecture as observed during a resting state acquisition. We demonstrate the use of our framework by comparing three tasks (visuo‐spatial working memory, reward processing, and stop signal task) available within a large cohort. Using task potency, we demonstrate that cognitive operations are supported by a set of common within‐network interactions, supplemented by connections between large‐scale networks in order to solve a specific task. HighlightsTask potency framework defines modulation of functional connectivity relative to resting state baseline.Task potency enables direct task comparison in terms of the amplitude of connectivity modulations.Task performance induces more within‐compared to between‐network modulations.Edges commonly modulated by multiple tasks are mostly within‐network.

Personality Profiles Are Associated with Functional Brain Networks Related to Cognition and Emotion

Personality factors as defined by the “five-factor model” are some of the most investigated characteristics that underlie various types of complex behavior. These are, however, often investigated as isolated traits that are conceptually independent, yet empirically are typically strongly related to each other. We apply Independent Component Analysis to these personality factors as measured by the NEO-FFI in 471 healthy subjects from the Human Connectome Project to investigate independent personality profiles that incorporate all five original factors. Subsequently we examine how these profiles are related to patterns of resting-state brain activity in specific networks-of-interest related to cognition and emotion. We find that a personality profile of contrasting openness and agreeableness is associated with engagement of a subcortical-medial prefrontal network and the dorsolateral prefrontal cortex. Likewise, a profile of contrasting extraversion and conscientiousness is associated with activity in the precuneus. This study shows a novel approach to investigating personality and how it is related to patterns of activity in the resting brain.

Quantitative Intensity Harmonization of Dopamine Transporter SPECT Images Using Gamma Mixture Models

PurposeDifferences in site, device, and/or settings may cause large variations in the intensity profile of dopamine transporter (DAT) single-photon emission computed tomography (SPECT) images. However, the current standard to evaluate these images, the striatal binding ratio (SBR), does not efficiently account for this heterogeneity and the assessment can be unequivalent across distinct acquisition pipelines. In this work, we present a voxel-based automated approach to intensity normalize such type of data that improves on cross-session interpretation.ProceduresThe normalization method consists of a reparametrization of the voxel values based on the cumulative density function (CDF) of a Gamma distribution modeling the specific region intensity. The harmonization ability was tested in 1342 SPECT images from the PPMI repository, acquired with 7 distinct gamma camera models and at 24 different sites. We compared the striatal quantification across distinct cameras for raw intensities, SBR values, and after applying the Gamma CDF (GDCF) harmonization. As a proof-of-concept, we evaluated the impact of GCDF normalization in a classification task between controls and Parkinson disease patients.ResultsRaw striatal intensities and SBR values presented significant differences across distinct camera models. We demonstrate that GCDF normalization efficiently alleviated these differences in striatal quantification and with values constrained to a fixed interval [0, 1]. Also, our method allowed a fully automated image assessment that provided maximal classification ability, given by an area under the curve (AUC) of AUC = 0.94 when used mean regional variables and AUC = 0.98 when used voxel-based variables.ConclusionThe GCDF normalization method is useful to standardize the intensity of DAT SPECT images in an automated fashion and enables the development of unbiased algorithms using multicenter datasets. This method may constitute a key pre-processing step in the analysis of this type of images.