2nd Cambridge Representational Similarity Analysis and Advanced Computational Methods Workshop (RSA2018)

After the successful RSA2015 Workshop (http://www.mrc-cbu.cam.ac.uk/rsa2015/) and the seminal paper on RSA Toolbox (Nili et al., 2014), there has been a notable increase in users of the toolbox, cited over hundred times since then.

Representational similarity analysis or RSA is a multivariate method for neuroimaging data based on comparing similarity in neural representations. It was initially applied to functional MRI data and led to a series of new discoveries on neuronal coding in visual and auditory systems in human and high profile publications. In the last few years, RSA has also been extended to other domains such as direct cell recording in nonhuman primates and other time series data including human MEG and EEG.

A number of new image processing and statistical methods have been developed in order to address the challenges introduced by time series data, such as controlling for multiple comparisons in spatiotemporal data. In addition, the ability of RSA in relating representations implemented in different systems has been exploited. During the advances of deep learning and other artificial intelligent techniques, computational models became powerful tools to understand human cognition and the brain. Thus, RSA has been applied to relate machine states/representations implemented in computational models with human brain states/representations implemented by neurons. This work has provided unprecedented knowledge about how both systems works.

In this Workshop, we will provide an in-depth introduction and advanced discussion of the current state-of-the-art in RSA and its tool supports. It will also cover the full range of existing and novel methods from data-driven visualization to inferential model comparisons, basic and advanced methods including recent methodological developments, shorter talks on recent neuroscientific and clinical applications and hands-on practicals.

This year, we will hold the event jointly with CMIH (Centre for mathematical imaging in healthcare) https://www.cmih.maths.cam.ac.uk, which kindly provided financial support. Speakers from CMIH will also present in this event expanding the content of this event to advanced multivariate, multimodal and multidimensional data analysis in general. At the end of this workshop, there will be an Open Discussion about how to further develop the RSA toolbox and other advanced computational methods in imaging, the challenges and opportunities to and form collaborations between research groups within Cambridge departments and more broadly.

The workshop will be suitable for participants from different disciplines at all levels of research experience. The methods are applicable to representations throughout the brain (e.g. vision, audition, memory, attention, emotion, language, and motor control) and to all modalities of brain-activity measurement. Most of the applications discussed will be in human neuroimaging (in particular, fMRI, EEG and MEG) in the fields of both cognitive and clinical neuroscience.

Preliminary speakers:

Mick Rugg, Johan Carlin, Cai Wingfield, Chris Petkov, Rik Henson, Carolin Sievers, Ed Bullmore, Hamed Nili, Alex Clarke, Tomas Cope, Li Su, Tim Kietzmann, Angelica Aviles-Rivero, Anders Hansen, Saber Sami, Alexandra Woolgar


Registration is free for this event however all delegates must register online (maximum 90 delegates). Delegate registration deadline 15th May 2018. Please follow the link (http://www.mrc-cbu.cam.ac.uk/internal-conferences/rsa2018/).