65ª Reunião Anual da SBPC

C. Ciências Biológicas - 12. Neurociências e Comportamento - 1. Neurociências e Comportamento

ENDOGENOUS MODULATION OF AFFILIATIVE EMOTION USING REAL-TIME FMRI NEUROFEEDBACK

Julie Hellen Weingartner - Universidade Federal do Rio de Janeiro / Instituto D’Or de Pesquisa e Ensino Patricia Pinheiro Bado - Universidade Federal do Rio de Janeiro / Instituto D’Or de Pesquisa e Ensino Jorge Moll Neto - Dr./Orientador- Instituto D’Or de Pesquisa e Ensino / UFRJ

INTRODUÇÃO:

Since first demonstrations that individuals can learn to control their own brain waves and associated psychological states (i.e.,EEG neurofeedback), several advances have emerged. Recently, the development of sophisticated brain decoding techniques and fMRI techniques have provided increased spatial and temporal accuracy, allowing real-time neurofeedback (fMRI-NFB) from specific cortical and subcortical regions. fMRI-NFB has been used to improve motor abilities, language, pain and basic emotions. In humans, the harmonic coexistence in a social group directly depends on mechanisms underlying cooperation/altruism, the essence of which is the human ability to experience affiliative emotion. We raised the hypothesis that neurofeedback provided according to each participant’s own brain activation pattern would allow individuals to gain voluntary control over the brain regions associated with these feelings. We also expected that neurofeedback would be capable of improving participants ability to attain specific affiliation-related brain activity. In a near future, neurofeedback may allow individuals to gain control or help in modulating deep brain regions in an attempt to increase adaptive emotional states.

OBJETIVO DO TRABALHO:

The objective of this work is to investigate the use of “brain decoding” algorithms on functional magnetic resonance imaging (fMRI) data to enable participants control their own brain activity (“neurofeedback”) over brain regions associated with affiliative emotion (more specifically, the feeling of tenderness).

MÉTODOS:

Participants (n=24, 14 females; age=24.5±3.4) were asked to recall real-life situations according to three emotional conditions: tenderness, self-pride and neutral. Participants’ instruction was to re-experience one specific emotion as intensely as possible while being scanned by the fMRI. In the first (training) acquisition, they were cued to each condition block, and evoked the respective scenario. The fMRI brain acquired data was used by an in-house software to train a multivariate classifier algorithm (Support Vector Machine, SVM) with brain activation patterns in order to distinguish in real-time two emotional conditions (tenderness and self-pride). Next, they underwent three “classification” runs while receiving either a control (CTR group) or a real visual neurofeedback (NFB group). The control stimuli consisted of a randomly distorting ring (unrelated to brain activity), whereas the real neurofeedback consisted of a ring that varied in distortion according to the output of the brain state classifier. The circle’s surface levels of distortion reflected how precisely the algorithm classified the ongoing emotional state. A smaller degree of distortion corresponded to a more precise classification. Functional images were acquired with a 3T Philips scanner.

RESULTADOS E DISCUSSÃO:

The NFB group displayed a higher frequency of brain activation patterns associated with tenderness at the last classification run when compared to the first (mean difference=26.0,SD=32.1, p=.02.).This effect was significantly different from the CTR group (p=.002) which did not show improvement (mean difference=-12.3,SD=21.7,p=.08).This corroborates our hypothesis that participants in NFB group would be able to use brain decoding signals to enhance complex brain activation patterns related to affiliative emotions compared with the CTR group.To compare our SVM-based brain decoding findings with the conventional fMRI approach, we further investigated the brain signature associated with affiliative emotion using offline functional MRI data by performing a general linear model analysis.We found an increased rate of activity of NFB group on septohypothalamic region and frontopolar cortex when covariating the accuracy increase with BOLD (blood-oxygen-level-dependent) activity for Tenderness vs Self-pride conditions.This finding is compatible with functional imaging and lesion studies addressing the neural bases of affiliative experiences,in-group perception, pro-social moral emotions and altruistic decisions, which also show involvement of septohypolalamic or frontopolar regions.

CONCLUSÕES:

Until recently, the prospect of a device that could measure and, possibly stimulate affiliative emotion, remained within the realm of science fiction. In this study we evaluated the role of neurofeedback in enhancing participants’ ability to evoke brain patterns associated with affiliative emotion. Our results indicate that multivariate brain decoding methods can be used to distinguish distributed patterns of brain activation associated with affiliative emotion from those associated with feelings of pride, and that this information can be conveyed using visual feedback signs that help participants boost consistency of brain state associated with tenderness feeling. These findings represent an important step in the quest for the development of brain-machine interfaces aiming to provide objective ways of endogenous modulation of complex affective states such as feelings of tenderness or interpersonal warmth, which tend to foster affiliation, trust, helping and care for others.

Palavras-chave: Affiliative emotion, Neurofeedback, fMRI.