Class imbalance should not throw you off balance: Choosing the right classifiers and performance metrics for brain decoding with imbalanced data.

Machine learning (ML) is increasingly used in cognitive, computational and clinical neuroscience. The reliable and efficient application of ML requires a sound understanding of its subtleties and limitations. Training ML models on datasets with imbalanced classes is a particularly common problem, and it can have severe consequences if not adequately addressed. With the neuroscience ML user in mind, this paper provides a didactic assessment of the class imbalance problem and illustrates its impact through systematic manipulation of data imbalance ratios in (i) simulated data and (ii) brain data recorded with electroencephalography (EEG), magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI). Our results illustrate how the widely-used Accuracy (Acc) metric, which measures the overall proportion of successful predictions, yields misleadingly high performances, as class imbalance increases. Because Acc weights the per-class ratios of correct predictions proportionally to class size, it largely disregards the performance on the minority class. A binary classification model that learns to systematically vote for the majority class will yield an artificially high decoding accuracy that directly reflects the imbalance between the two classes, rather than any genuine generalizable ability to discriminate between them. We show that other evaluation metrics such as the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC), and the less common Balanced Accuracy (BAcc) metric - defined as the arithmetic mean between sensitivity and specificity, provide more reliable performance evaluations for imbalanced data. Our findings also highlight the robustness of Random Forest (RF), and the benefits of using stratified cross-validation and hyperprameter optimization to tackle data imbalance. Critically, for neuroscience ML applications that seek to minimize overall classification error, we recommend the routine use of BAcc, which in the specific case of balanced data is equivalent to using standard Acc, and readily extends to multi-class settings. Importantly, we present a list of recommendations for dealing with imbalanced data, as well as open-source code to allow the neuroscience community to replicate and extend our observations and explore alternative approaches to coping with imbalanced data.

Dancing to "groovy" music enhances the experience of flow.

We investigated whether dancing influences the emotional response to music, compared to when music is listened to in the absence of movement. Forty participants without previous dance training listened to "groovy" and "nongroovy" music excerpts while either dancing or refraining from movement. Participants were also tested while imitating their own dance movements, but in the absence of music as a control condition. Emotion ratings and ratings of flow were collected following each condition. Dance movements were recorded using motion capture. We found that the state of flow was increased specifically during spontaneous dance to groovy excerpts, compared with both still listening and motor imitation. Emotions in the realms of vitality (such as joy and power) and sublimity (such as wonder and nostalgia) were evoked by music in general, whether participants moved or not. Significant correlations were found between the emotional and flow responses to music and whole-body acceleration profiles. Thus, the results highlight a distinct state of flow when dancing, which may be of use to promote well-being and to address certain clinical conditions.

Enhancement of Pleasure during Spontaneous Dance.

Dancing emphasizes the motor expression of emotional experiences. The bodily expression of emotions can modulate the subjective experience of emotions, as when adopting emotion-specific postures and faces. Thus, dancing potentially offers a ground for emotional coping through emotional enhancement and regulation. Here we investigated the emotional responses to music in individuals without any prior dance training while they either freely danced or refrained from movement. Participants were also tested while imitating their own dance movements but in the absence of music as a control condition. Emotional ratings and cardio-respiratory measures were collected following each condition. Dance movements were recorded using motion capture. We found that emotional valence was increased specifically during spontaneous dance of groovy excerpts, compared to both still listening and motor imitation. Furthermore, parasympathetic-related heart rate variability (HRV) increased during dance compared to motor imitation. Nevertheless, subjective and physiological arousal increased during movement production, regardless of whether participants were dancing or imitating. Significant correlations were found between inter-individual differences in the emotions experienced during dance and whole-body acceleration profiles. The combination of movement and music during dance results in a distinct state characterized by acutely heightened pleasure, which is of potential interest for the use of dance in therapeutic settings.