"Workshop on Mechanism of Brain and Mind" successfully concluded
We co-organized “Workshop on Mechanism of Brain and Mind” in Rusutsu Hokkaido on January 9-11, 2018. Abstracts and References are available from the Workshop page.
The theme of the winter workshop was “Body control and self-representation”. Investigation of how the brain generates behavior by modeling itself and controlling the body is the approach that is complementary to studies of perception in the pursuit of the common goal, i.e. understanding the nature of consciousness, self-awareness and other cognitive processes. The workshop enabled experienced scientists to present results of their decades-long research projects as well as fresh academic recruits to demonstrate preliminary results of their theses. The presented studies demonstrated a multitude of approaches to understanding how body control and self-representation are implemented. These range from neuroimaging studies and computational models of the neural mechanisms to behavioral studies and models on the level of individuals and even societies.
The notions of hierarchy and prediction have been crucial in studying the neural mechanisms controlling the body and underlying self-awareness. For example, Professor Jun Tani presented data from neurorobotic experiments in which a recurrent neural network implemented a recognition process in a hierarchical bottom-up manner from fast to slow representations. The timescale hierarchy was achieved in the networks built on predictive coding principles, such as the propagation of prediction and prediction error signals. The author believes that the interaction between bottom-up recognition and top-down intentions may account for the emergence of conscious phenomena.
Notably, predictive coding finds its reflection in the sense of agency, i.e. attributing actions distinctly to oneself or the external world. Professor Hiroaki Gomi conducted behavioral experiments studying the interaction between visual and motor motions in a target reaching task. The data suggest that the lack of motor-dependent sensory predictions can be the cause of visual illusions such as Duncker illusion.
Researcher Atsushi Yokoi used fMRI experiments to directly investigate neural hierarchical representations of simple and complex motor sequences. The presented work indicated possible coexistence of both hierarchical and non-hierarchical representations in primary and secondary motor areas of the cerebral cortex.
In the behavioral study and the computational model by researcher Ken Takiyama, the errors of future actions were predicted, which allowed to explain and reproduce multiple motor-learning-related phenomena.
Technologies such as virtual reality have enabled researchers to better approach the questions related to self-representation, including social interactions. Researcher Antonia Hamilton presented her work in which she used virtual reality and near-infrared spectroscopy to probe the mechanisms of imitation. Researcher Tatsuya Kameda used computer simulations, anthropological and behavioral data to argue that egalitarian behavior can be an adaptive device to reduce statistical risks involved in resource acquisition.
Metacognition was tackled in the experiments by researcher Aurelio Cortese, who used real-time fMRI to deliver neurofeedback and manipulate neural correlates of confidence. The participants could implicitly access the spontaneous neural activity that determined the task, thus maximizing their reward. In addition, learning resulted in increased connectivity between basal ganglia and prefrontal regions.
The molecular mechanisms of memory were addressed in the presentation by researcher Hiroshi Nomura. Their studies showed that the weak long-term memory engrams can be reactivated through activation of histamine, thus restoring access to forgotten memories. This was achieved by administering inverse agonists of histamine H3 receptors that downregulate histamine synthesis and release.
Researcher Masafumi Oizumi used Integrated Information Theory to make predictions about the boundary problem of consciousness. Based on the theory, consciousness results from the locally maximized integration of information. The author discussed theoretical predictions and possible ways to test them experimentally.
Finally, the poster presentations allowed fresh research projects to be presented, discussed and constructively criticized. A wide range of projects from computational models to neuroimaging and behavioral studies was presented.
Together, the presented findings and preliminary results will inspire further scientific progress towards understanding the mechanisms of brain and mind.
Sergey Zobnin (Okinawa Institute of Science & Technology Graduate University, Neural Computation Unit)
