Brain Imaging Methods in Social and Affective Neuroscience: A Machine Learning Perspective

Machine learning (ML) is a subarea of artificial intelligence which uses the induction approach to learn based on previous experiences and make conclusions about new inputs (Mitchell, Machine learning. McGraw Hill, 1997). In the last decades, the use of ML approaches to analyze neuroimaging data has attracted widening attention (Pereira et al., Neuroimage 45(1):S199–S209, 2009; Lemm et al., Neuroimage 56(2):387–399, 2011). Particularly interesting recent applications to affective and social neuroscience include affective state decoding, exploring potential biomarkers of neurological and psychiatric disorders, predicting treatment response, and developing real-time neurofeedback and brain-computer interface protocols. In this chapter, we review the bases of the most common neuroimaging techniques, the basic concepts of ML, and how it can be applied to neuroimaging data. We also describe some recent examples of applications of ML-based analysis of neuroimaging data to social and affective neuroscience issues. Finally, we discuss the main ethical aspects and future perspectives for these emerging approaches.