Seminars

Understanding Human ability to learn novel concepts from just a few sensory experiences is a fundamental problem in cognitive neuroscience. I will describe a recent work with Ben Sorcher and Surya Ganguli (PNAS, October 2022) in which we propose a simple, biologically plausible, and mathematically tractable neural mechanism for few-shot learning of naturalistic concepts. We posit that the concepts that can be learned from few examples are defined by tightly circumscribed manifolds in the neural firing-rate space of higher-order sensory areas. Discrimination between novel concepts is performed by downstream neurons implementing ‘prototype’ decision rule, in which a test example is classified according to the nearest prototype constructed from the few training examples. We show that prototype few-shot learning achieves high few-shot learning accuracy on natural visual concepts using both macaque inferotemporal cortex representations and deep neural network (DNN) models of these representations. We develop a mathematical theory that links few-shot learning to the geometric properties of the neural concept manifolds and demonstrate its agreement with our numerical simulations across different DNNs as well as different layers. Intriguingly, we observe striking mismatches between the geometry of manifolds in intermediate stages of the primate visual pathway and in trained DNNs. Finally, we show that linguistic descriptors of visual concepts can be used to discriminate images belonging to novel concepts, without any prior visual experience of these concepts (a task known as ‘zero-shot’ learning), indicated a remarkable alignment of manifold representations of concepts in visual and language modalities. I will discuss ongoing effort to extend this work to other high level cognitive tasks.

To efficiently process sensory information, the brain relies on statistical regularities in the input. While generally improving the reliability of sensory estimates, this strategy also induces perceptual illusions that help reveal the underlying computational principles. Focusing on auditory and visual perception, in my talk I will describe how the brain exploits statistical regularities within and across the senses for the perception space, time and multisensory integration. In particular, I will show how results from a series of psychophysical experiments can be interpreted in the light of Bayesian Decision Theory, and I will demonstrate how such canonical computations can be implemented into simple and biologically plausible neural circuits. Finally, I will show how such principles of sensory information processing can be leveraged in virtual and augmented reality to overcome display limitations and expand human perception.

Several lines of research have demonstrated that internally generated sensory experience - such as during memory, dreaming and mental imagery - activates similar neural representations as externally triggered perception. This overlap raises a fundamental challenge: how is the brain able to keep apart signals reflecting imagination and reality? In a series of online psychophysics experiments combined with computational modelling, we investigated to what extent imagination and perception are confused when the same content is simultaneously imagined and perceived. We found that simultaneous congruent mental imagery consistently led to an increase in perceptual presence responses, and that congruent perceptual presence responses were in turn associated with a more vivid imagery experience. Our findings can be best explained by a simple signal detection model in which imagined and perceived signals are added together. Perceptual reality monitoring can then easily be implemented by evaluating whether this intermixed signal is strong or vivid enough to pass a ‘reality threshold’. Our model suggests that, in contrast to self-generated sensory changes during movement, our brain does not discount self-generated sensory signals during mental imagery. This has profound implications for our understanding of reality monitoring and perception in general.

Selection is the hallmark of attention: processing improves for attended items but is relatively impaired for unattended items. It is well known that visual spatial attention changes sensory signals and perception in this selective fashion. In the work I will present, we asked whether and how attentional selection happens across time. First, our experiments revealed that voluntary temporal attention (attention to specific points in time) is selective, resulting in perceptual tradeoffs across time. Second, we measured small eye movements called microsaccades and found that directing voluntary temporal attention increases the stability of the eyes in anticipation of an attended stimulus. Third, we developed a computational model of dynamic attention, which proposes specific mechanisms underlying temporal attention and its selectivity. Lastly, I will mention how we are testing predictions of the model with MEG. Altogether, this research shows how precisely timed voluntary attention helps manage inherent limits in visual processing across short time intervals, advancing our understanding of attention as a dynamic process.

When we remember a stimulus we rarely maintain a full fidelity representation of the observed item. Our working memory instead maintains a mixture of the observed feature values and categorical/gist information. I will discuss evidence from computational models supporting a mix of categorical and fine-detail information in working memory. Having established the need for two memory formats in working memory, I will discuss whether categorical and fine-detailed information for a stimulus are represented separately or as a single unified representation. Computational models of these two potential cognitive structures make differing predictions about the pattern of responses in visual working memory recall tests. The present study required participants to remember the orientation of stimuli for later reproduction. The pattern of responses are used to test the competing representational structures and to quantify the relative amount of fine-detailed and categorical information maintained. The effects of set size, encoding time, serial order, and response order on memory precision, categorical information, and guessing rates are also explored. (This is a 60 min talk).

The talk will give a tour of Guided Search 6.0 (GS6), the latest evolution of the Guided Search model of visual search. Part 1 describes The Mechanics of Search. Because we cannot recognize more than a few items at a time, selective attention is used to prioritize items for processing. Selective attention to an item allows its features to be bound together into a representation that can be matched to a target template in memory or rejected as a distractor. The binding and recognition of an attended object is modeled as a diffusion process taking > 150 msec/item. Since selection occurs more frequently than that, it follows that multiple items are undergoing recognition at the same time, though asynchronously, making GS6 a hybrid serial and parallel model. If a target is not found, search terminates when an accumulating quitting signal reaches a threshold. Part 2 elaborates on the five sources of Guidance that are combined into a spatial “priority map” to guide the deployment of attention (hence “guided search”). These are (1) top-down and (2) bottom-up feature guidance, (3) prior history (e.g. priming), (4) reward, and (5) scene syntax and semantics. Finally, in Part 3, we will consider the internal representation of what we are searching for; what is often called “the search template”. That search template is really two templates: a guiding template (probably in working memory) and a target template (in long term memory). Put these pieces together and you have GS6.

Human vision is full of puzzles. Observers can grasp the essence of a scene in an instant, yet when probed for details they are at a loss. People have trouble finding their keys, yet they may be quite visible once found. How does one explain this combination of marvelous successes with quirky failures? I will describe our attempts to develop a unifying theory that brings a satisfying order to multiple phenomena. One key is to understand peripheral vision. A visual system cannot process everything with full fidelity, and therefore must lose some information. Peripheral vision must condense a mass of information into a succinct representation that nonetheless carries the information needed for vision at a glance. We have proposed that the visual system deals with limited capacity in part by representing its input in terms of a rich set of local image statistics, where the local regions grow — and the representation becomes less precise — with distance from fixation. This scheme trades off computation of sophisticated image features at the expense of spatial localization of those features. What are the implications of such an encoding scheme? Critical to our understanding has been the use of methodologies for visualizing the equivalence classes of the model. These visualizations allow one to quickly see that many of the puzzles of human vision may arise from a single encoding mechanism. They have suggested new experiments and predicted unexpected phenomena. Furthermore, visualization of the equivalence classes has facilitated the generation of testable model predictions, allowing us to study the effects of this relatively low-level encoding on a wide range of higher-level tasks. Peripheral vision helps explain many of the puzzles of vision, but some remain. By examining the phenomena that cannot be explained by peripheral vision, we gain insight into the nature of additional capacity limits in vision. In particular, I will suggest that decision processes face general-purpose limits on the complexity of the tasks they can perform at a given time.

Our research investigates how processes of attention, visual working memory (VWM), and decision-making combine to translate perception into action. Within this framework, the role of VWM is to form stable representations of transient stimulus events that allow them to be identified by a decision process, which we model as a diffusion process. In psychophysical tasks, we find the capacity of VWM is well defined by a sample-size model, which attributes changes in VWM precision with set-size to differences in the number evidence samples recruited to represent stimuli. In the first part of the talk, I review evidence for the sample-size model and highlight the model's strengths: It provides a parameter-free characterization of the set-size effect; it has plausible neural and cognitive interpretations; an attention-weighted version of the model accounts for the power-law of VWM, and it accounts for the selective updating of VWM in multiple-look experiments. In the second part of the talk, I provide a characterization of the theoretical relationship between two-choice and continuous-outcome decision tasks using the circular diffusion model, highlighting their common features. I describe recent work characterizing the joint distributions of decision outcomes and response times in continuous-outcome tasks using the circular diffusion model and show that the model can clearly distinguish variable-precision and simple mixture models of the evidence entering the decision process. The ability to distinguish these kinds of processes is central to current VWM studies.

I will begin by describing recent research taking a new, model-based approach to perceptual development. This approach uncovers fundamental changes in information processing underlying the protracted development of perception, action, and decision-making in childhood. For example, integration of multiple sensory estimates via reliability-weighted averaging – widely used by adults to improve perception – is often not seen until surprisingly late into childhood, as assessed by both behaviour and neural representations. This approach forms the basis for a newer question: the scope for the nervous system to deploy useful computations (e.g. reliability-weighted averaging) to optimise perception and action using newly-learned sensory signals provided by technology. Our initial model system is augmenting visual depth perception with devices translating distance into auditory or vibro-tactile signals. This problem has immediate applications to people with partial vision loss, but the broader question concerns our scope to use technology to tune in to any signal not available to our native biological receptors. I will describe initial progress on this problem, and our approach to operationalising what it might mean to adopt a new signal comparably to a native sense. This will include testing for its integration (weighted averaging) alongside the native senses, assessing the level at which this integration happens in the brain, and measuring the degree of ‘automaticity’ with which new signals are used, compared with native perception.

The talk will give a tour of Guided Search 6.0 (GS6), the latest evolution of Guided Search. Part 1 describes The Mechanics of Search. Because we cannot recognize more than a few items at a time, selective attention is used to prioritize items for processing. Selective attention to an item allows its features to be bound together into a representation that can be matched to a target template in memory or rejected as a distractor. The binding and recognition of an attended object is modeled as a diffusion process taking > 150 msec/item. Since selection occurs more frequently than that, it follows that multiple items are undergoing recognition at the same time, though asynchronously, making GS6 a hybrid serial and parallel model. If a target is not found, search terminates when an accumulating quitting signal reaches a threshold. Part 2 elaborates on the five sources of Guidance that are combined into a spatial “priority map” to guide the deployment of attention (hence “guided search”). These are (1) top-down and (2) bottom-up feature guidance, (3) prior history (e.g. priming), (4) reward, and (5) scene syntax and semantics. In GS6, the priority map is a dynamic attentional landscape that evolves over the course of search. In part, this is because the visual field is inhomogeneous. Part 3: That inhomogeneity imposes spatial constraints on search that described by three types of “functional visual field” (FVFs): (1) a resolution FVF, (2) an FVF governing exploratory eye movements, and (3) an FVF governing covert deployments of attention. Finally, in Part 4, we will consider that the internal representation of the search target, the “search template” is really two templates: a guiding template and a target template. Put these pieces together and you have GS6.