The 11th Australasian Workshop on Neuro-Engineering and Computational Neuroscience will be a three-day conference hosted at Sydney University from 27th-29th November, 2018.
Propagating waves in visual cortex and their computational role
Propagating waves have been observed in various in vivo and in vitro preparations but their role still remains obscure. Propagating waves were also seen in physiological conditions, such as in primary and secondary visual cortical areas of awake monkey (Muller et al., Nature Commun. 2014), suggesting they may play a role in vision. Here, using voltage-sensitive dye imaging in awake monkey V1, we show that colliding propagating waves always sum sublinearly and mediate a suppressive effect. By using a probabilistic decoder, we show that this suppression enables to dis-ambiguate stimuli. These features were also captured by computational models, using a mean-field approach. The model suggests that the suppressive effect depends on two ingredients, the fact that inhibition has a higher gain than excitation, and the fact that they combine via conductance-based interactions. In conclusion, we show a plausible role for propagating waves in the processing of visual information.
Brain Songs: Discovering the relevant timescale and richness of repertoire of the human brain
A key unresolved problem in neuroscience is to determine the relevant timescale for understanding spatiotemporal dynamics across the whole brain. Yet, a significant problem has been the lack of whole-brain neuroimaging data at many different timescales. Much recent focus has been on discovering spatial correlation brain maps in functional magnetic resonance imaging (fMRI) data. The timescale of resting state networks has been found to be ultraslow (below 0.1Hz), which makes fMRI an appropriate tool for discovery for this timescale. Some evidence has come from other neuroimaging modalities such as MEG and EEG (albeit with more limited spatial resolution) suggesting that much faster timescales may be equally or more relevant for discovering spatiotemporal structure beyond static spatial maps derived from averaged signals. We propose a novel way to resolve these problems by using whole-brain modelling of fMRI signals allowing us to generate whole-brain neural dynamical activity at the millisecond scale. This method allows us to study the different timescales through binning, from milliseconds to seconds, the output of the model. These timescales can then be investigated using a novel method (poetically named brain songs) to extract the spacetime motifs at a given timescale. Using independent measures of entropy and hierarchy to characterize the richness of the dynamical repertoire, we show that both methods find a similar optimum at a timescale of around 200 ms in resting state and in task data. Significantly, we confirm the timescale in MEG data without using whole-brain modelling. This result is consistent with the emerging evidence from other studies for the importance of this timescale for allowing the optimal repertoire for brain dynamics.
Peer into your brain via contiguous scale data
Using genetics, imaging, behaviors and environmental data, together with advanced machine learning algorithms, we are able to assess the abnormalities of a patient brain and the cognitive activities of a healthy brain. In this talk we will concentrate on depression and schizophrenia. In depression, using the BWAS method we developed, we identified that an unbalanced reward and punishment sensitivity is the possible roots of the disease, as further confirmed in TMS treatments. As the key location of emotion, we found that OFC plays the central role in sleeping, drinking and smoking, although in drinking and smoking the functional connectivity changed in opposite directions as in depression. In schizophrenia, using the IDSCN method, it is found that the link between putamen and hippocampus is abnormal, with its genetic variant being SNP rs13107325. In summary, the reward and nonreward part of the brain, which is the key area for many mental disorders, should be an integrated part of an artificial intelligent system.
Abstract submission is NOW CLOSED. Abstracts selected for an oral presentation will be notified early in October, with preference given to speakers who have pre-registered.
General registration is open until 9th November: $300 + GST (students), $540 + GST (general) [includes social mixer and conference dinner].
This first day, themed New Directions in Brain Modeling and Analysis: Dynamics, Structure, and Function aims to present new and general methods for quantitative neuroscience. Presentations will introduce new tools and techniques and may include interactive software demonstrations.
Being just a week away from the end of residential contracts, there will be many rooms available in college accomodation on campus, at affordable rates. You may contact Sydney University accomodation services to inquire directly via email. You may also investigate options through University Stays
We have negotiated discount rates for the following two hotels:
Mercure Sydney (4 star)
818-820 George Street, Chippendale.
Tel. +61 2 9217 6666
1.6km from conference venue (20 min walk or 12 min by bus)
Rate for conference delegates: $230/night (optional: +$25 breakfast)
Ibis World Square (3.5 star)
382–384 Pitt Street, Sydney.
Tel. +61 2 8627 3111
2.5km from conference venue (31 min walk or 18 min by bus)
Rate for conference delegates: $200/night (optional: +$25 breakfast)
You can access these reduced rates by setting your DESTINATION as either "MERCURE SYDNEY" or "IBIS SYDNEY WORLD SQUARE" when you book here