xEarly brain functional studies, based on MRI, PET or EEG, focused on univariate analyses, in which the activity of each region is processed independently from each other. Nowadays, multivariate machine learning techniques have been developed to model complex, sparse neuronal populations. This course will provide an introduction to new methods and cutting-edge machine-learning techniques in the neuroimaging field by exploring multivariate statistical modeling of brain-activity data and computational modeling of brain information processing. Specifically, the course focuses on machine learning decoding and encoding perspectives in fMRI and novel methods (e.g., Representational Similarity Analysis) to explore and analyze brain data. A comprehensive review of model validation and statistical inference is provided. In addition, hardware and software implementation recently allowed to combine different neural measures with different spatial and temporal resolutions within the same experimental session. The course also discusses the transdisciplinary approach combining different neuroimaging techniques in unique methodological frameworks and the advent of ultrahigh field neuroimaging.