10:30 Arrival and coffee (Physics B floor balcony)
11:00 Marcus Kaiser (University of Nottingham)
Title: Using connectome-based computational models to inform brain stimulation interventions
Abstract: The complete set of connections in the brain is called our connectome. Over the last 20 years we have found out more about how this network is organised and how this organisation is linked to brain function. For example, highly-connected brain regions (hubs) play critical roles in information processing and are involved in many brain diseases. I will outline how networks change for a range of brain disorders, from networks that produce seizures for epilepsy to networks that produce hallucinations in certain types of dementia. Given these changes, can we alter the structure of these networks and thereby improve cognition in patients?
Brain stimulation is an option to achieve this and has been proposed as an alternative treatment to pharmaceutical drugs with a potential to reduce side effects and improve cognitive function. I will outline how computational models based on brain connectivity information can help to identify network targets and to find personalised stimulation protocols. In particular, I will highlight how focused ultrasound, a novel non-invasive technology for brain stimulation, can directly target deep-brain structures involved in emotion and memory processing opening up a way to new interventions. More information can also be found in my book ‘Changing Connectomes’ (MIT Press, 2020; https://mitpress.mit.edu/books/changing-connectomes).
11:45 Yulia Timofeeva (University of Warwick)
Title: Virtual presynaptic nerve terminal
Abstract: Synaptic transmission provides the basis for neuronal communication. When an action-potential propagates through the axonal arbour, it activates voltage-gated Ca2+ channels located in the vicinity of release-ready synaptic vesicles docked at the presynaptic active zone. Ca2+ ions enter the presynaptic terminal and activate the vesicular Ca2+ sensor, thereby triggering neurotransmitter release. This whole process occurs on a timescale of a few milliseconds. In addition to fast, synchronous release, which keeps pace with action potentials, many synapses also exhibit delayed asynchronous release that persists for tens to hundreds of milliseconds. In this talk I will demonstrate how experimentally constrained computational modelling of underlying biological processes can complement laboratory studies (using electrophysiology and imaging techniques) and provide insights into the mechanisms of synaptic transmission.
12:30 Lunch and posters (Mathematical Sciences Atrium)
13:45 Mark van Rossum (University of Nottingham)
Title: Energy efficient learning within neural networks
Abstract: The brain is one of the most energy intense organs. Some of this energy is used for neural information processing, however, fruitfly experiments have shown that also synaptic plasticity is metabolically costly.
First we will present estimates of this cost, introduce a general model of this cost, and compare it to costs in computers Next, we turn to a supervised artificial network setting and explore a number of strategies that can save energy need for plasticity, either by modifying the cost function, by restricting plasticity, or by using less costly transient forms of plasticity.
Finally, we will discuss adaptive strategies and possible relevance for computer hardware.
14:30 Lei Zhang (University of Birmingham)
Title: Modelling flexible behaviour in autism spectrum disorders through the lens of simple reinforcement learning
Abstract: One of the core symptoms of autism is repetitive behavior, which is likely to be underpinned by cognitive inflexibility, i.e., the deficit in responding to changes in the environment. I will present how to measure flexible behavior with a probabilistic learning task, and how to quantify the suboptimal learning in autism using simple reinforcement learning models, showcasing a newly emerged field - computational psychiatry.
15:00 Tea and coffee (Physics B floor balcony)
15:30 Farzaneh Darki (University of Exeter)
Title: Hierarchical processing underpins competition in tactile perceptual bistability
Abstract: Ambiguous sensory information can lead to spontaneous alternations
between perceptual states, known as perceptual rivalry. Earlier studies
with tactile stimuli identified bistability for tactile sensation [1-3]. We
recently proposed a simple form of tactile rivalry where stimuli consisted
of antiphase sequences of high and lowintensity pulses delivered to the
right and left index fingers (Fig 1A) . Stimuli were perceived as either
one simultaneous pattern of vibration on both hands (SIM), or patterns of
vibration that jumped from one hand to the other, giving a sensation of
apparent movement (AM). Our quantitative analysis of alternation times
allowed for direct comparisons with data from other sensory modalities.
This study addresses a need for tactile rivalry model that accounts for
well-established results on dynamics of perceptual alternations and that is
compatible with the structure of the somatosensory system (Fig 1B). The
presented model includes hierarchical processing (Fig 1C); a first stage
resolves perceptual competition, leading to perceptual alternations; and a
second stage encodes perceptual interpretations. The first and the second
stages of model could be located at the secondary somatosensory cortex
(area S2), or in higher areas driven by S2 .
As stimuli are antiphase and given the symmetry in the first stage, our
bioinspired tactile rivalry model can be simplified. Bifurcation analysis of
the simple model was used to tune parameters of the first stage to
operate within an oscillatory regime  and of the second stage to
operate within a range where direct transitions between SIM and AM occur
. Other parameters were tuned using a genetic algorithm to minimise
the differences between the experimental and computational mean
dominance. Beside capturing dynamical features of perceptual
interpretations in tactile rivalry, this model can produce general features
of perceptual rivalry including Levelt's proposition, short-tailed skewness
of reversal time distributions (Fig 1D-E) and the ratio of distribution
moments. This approach could be extended to explore positive sequential
correlation of dominance periods and a scaling property. The presented
modelling work leads to experimentally testable predictions and the same hierarchical model could generalise to account for perceptual bistability in
visual and auditory domains.
Figure 1 A) Tactile stimuli and percepts. B) Afferent fibres cross over and
project to thalamic nuclei on opposite side, then project to cerebral
cortex. C) Bioinspired tactile rivalry model. D) Levelt’s proposition, Input
nonlinearity. E) Log-normal and gamma distribution for normalized
 Carter, O., Konkle, T., Wang, Q., Hayward, V., and Moore, C. (2008).
Tactile rivalry demonstrated with an ambiguous apparent-motion quartet.
Current Biology, 18(14):1050–1054
 Liaci, E., Bach, M., van Elst, L. T., Heinrich, S. P., and Kornmeier, J.
(2016). Ambiguity in tactile apparent motion perception. PLoS One, 11(5).
 Conrad, V., Vitello, M. P., and Noppeney, U. (2012). Interactions
between apparent motion rivalry in vision and touch. Psychological
 Darki, F. and Rankin, J. (2021). Perceptual rivalry with vibrotactile
stimuli. Attention, Perception, & Psychophysics, pages 1–12.
 Delhaye, B. P., Long, K. H., and Bensmaia, S. J. (2011). Neural basis of
touch and proprioception in primate cortex. Comprehensive Physiology,
 Levenstein, D., Buzsáki, G., and Rinzel, J. (2019). Nrem sleep in the
rodent neocortex and hippocampus reflects excitable dynamics. Nature
 Ferrario, A. and Rankin, J. (2021). Auditory streaming emerges from
fast excitation and slow delayed inhibition. The Journal of Mathematical
16:00 Carolina Feher da Silva (University of Nottingham)
Title: Model-free or muddled models in the two-stage task?
Abstract: A standard assumption in decision science is that low-effort model-free learning is automatic and continuously employed, while more complex model-based strategies are only used when the rewards they generate are worth the additional effort (i.e., opportunistically). I will present evidence from behavioural and neural data refuting this assumption in two ways. First, I will demonstrate that previous reports of combined model-free and model-based reward prediction errors in the ventral striatum are based on a flawed statistical model and most likely spurious. Analyses with an appropriate model yield no evidence of a model-free prediction error signal in the ventral striatum. Second, task instructions that lead to more correct model-based behaviour reduce rather than increase mental effort. This is inconsistent with cost-benefit arbitration about whether to add opportunistic model-based strategies to the fixed cost of automatic model-free learning. Together, these physiological and behavioural data suggest that model-free learning may not be automatic. Instead, humans can reduce mental effort by using a model-based strategy alone rather than arbitrating between multiple strategies. These results call for re-evaluation of the core assumptions in influential theories of learning and decision-making.
16:30 Michael Okun (University of Sheffield)
Title: Brain state transitions primarily impact the rate of slow-firing neurons
Abstract: The spontaneous firing rate of forebrain neurons spans at least five orders of magnitude (~0.001—100 spikes/s), yet across distinct brain states the rate of individual neurons typically does not change more than two-three-fold. Given this relatively fixed position of each neuron on the rate spectrum, are there consistent distinctions in the way fast-firing and slow-firing neurons (multiplicatively) modulate their rate across brain states? To address this question, we analysed electrophysiological neuronal population recordings from rodent cortex, hippocampus and thalamus across transitions between wakefulness and sleep, changes in the level of arousal during wakefulness, and following administration of psychoactive drugs.
Typically, a brain state transition both upregulated and downregulated large neuronal subpopulations. We therefore considered these two subpopulations separately. Consistently across all the considered brain areas and categories of brain states, we observed that for an upregulated subpopulation the bell-shaped distribution of its log-rates could get narrower but not wider. Conversely, the log-rate distribution of a downregulated subpopulation only got wider. This implies that in both upregulated and downregulated subpopulations, the rate of fast-firing neurons was modulated substantially less than of the slow-firing neurons. Further analysis of the modulation variance indicated that across fast-firing neurons weak modulation was almost universal, whereas across slow-firing neurons high variability in the strength of modulation existed. The subpopulation of ‘malleable’
slow-firing neurons was unique to each category of brain state transition. This empirical modulation structure could be recapitulated by a correlated bivariate log-gamma distribution, whose marginals have long left tails of slow-firing neurons, but not by a bivariate Gaussian.