Abstract
How does consciousness arise from neural activity? Several theories try to answer this question, each providing fundamentally different interpretations of empirical data. Accordingly, no agreed-upon, comprehensive account of the neural substrates of consciousness has been achieved. To date, overviews in the field have relied mostly on qualitative reviews of the results and were typically written from the standpoint of one theory. Thus, a systemic, quantitative and theory-free review of studies is necessary to characterize the state of the field regarding these theories. Here, we conducted a comprehensive, quantitative and theory-neutral review of neuroscientific studies of consciousness (379 papers, reporting 418 experiments) that interpreted their findings as supporting/challenging one of four leading theories of consciousness: Global Neuronal Workspace, Integrated Information Theory, Recurrent Processing Theory and Higher Order Theory. Then, to identify biases in the way these theories were studied, the distributions of experiments to parameters extracted from each experiment were compared. Notably, all data is available in an open-access website with interactive plotting tools, to allow other researchers to conduct additional queries and analysis of the data. We found that some methodological choices of researchers increase the probability of their findings supporting certain theories. Also, we found that the field generally suffers from a strong confirmatory-bias, and that the majority of studies post-hoc interpret their findings concerning the theories, rather than designed a-priori to test their critical predictions, and that. Finally, when all findings were collapsed together, a highly variable pattern of spatial and temporal findings emerged, that is not predicted by any theory. To overcome these biases, potentially converging towards an agreed upon account, cross-talk between the theories is needed, testing each other's predictions and integrating ideas, as opposed to gaining further affirmative results according to each theory’s predictions.