Research

Research in the Icelandic Vision Lab covers a wide field of topics. Here are some examples.


“Big Five” of Object Perception

What are the principal diagnostic dimensions on which objects differ visually from one another? Do these dimensions provide the foundation for the organizational principles of visual object perception, as evidenced by individual differences in behavior, special abilities and disabilities, and developmental trajectories? With the aid of convolutional neural networks trained on image categorization, we are estimating the dimensions on which individuals differ in object perception, both behaviorally and neurally.

People mainly involved in this project:

Inga María Ólafsdóttir
Heida Maria Sigurdardottir

Computational Modelling

Thousands of computational models of vision currently exist, however, most of them lack the critical aspects of visual salience modelling from a computational cognitive neuroscience perspective. Such a perspective suggests that the perfect salience model should be based on a strong theoretical foundation, model the neurobiological processes underlying visual saliency, use explicit computational tools to model them, and be generated by taking both spatial and temporal predictions of visual salience into account. This project attempts to improve spatiotopic salience models by using additional information gathered through several eyetracking experiments. Using a previously created version of a deep learning saliency model (Krasovskaya, Zhulikov & MacInnes, 2021), we use the collected and processed data to build an improved version of a generative neural network model of visual attention. This approach adds to our understanding of the neural correlates of visual system mechanisms, such as saccade planning and generation. The project is conducted in collaboration with the Vision Modelling Lab (VML), HSE University, Moscow.

People involved in this project:

Sofia Krasovskaya
Árni Kristjánsson
Joseph MacInnes (VML)


The Functional Field of View

The amount of visual information conveyed through the retina is limited by the size of the spatial area surrounding the fixation point. This area, called the Functional Field of View (FFV), prevents the visual system from being overloaded. The size of the FFV is determined by two types of factors that drive visual perception, namely, the retinotopic (physiological) and the functional (cognitive) factors. One hypothesis is that the FFV is highly adaptable and changes based on the difficulty of the given task (Young & Hulleman, 2013), shrinking in response to high task demands and expanding during the execution of easier tasks. We study the phenomenon in a retinotopic- native for the visual system – frame of reference by adding a gaze- contingent setup. This study is conducted in collaboration with the Vision Modelling Lab (VML), HSE University, Moscow.

People involved in this project:

Sofia Krasovskaya
Árni Kristjánsson
Joseph MacInnes (VML)


Microsaccadic Eye Movements in the Antisaccade Paradigm

Despite over 60 years of research, microsaccadic eye movements are still a hot topic for debate in the field of vision research. Microsaccades belong to the category of fixational micromovements and may be crucial for image stability on the retina. During the so-called antisaccade task, planning and execution are separate processes. Our current work is inspired by the hypothesis that microsaccade rates may be reduced prior to the execution of antisaccades as compared to regular saccades. We hope that this project would contribute to the existing literature and possibly improve our understanding of the mysterious eye movement. This study is conducted in collaboration with the Vision Modelling Lab (VML), HSE University, Moscow.

People involved in this project:

Sofia Krasovskaya
Árni Kristjánsson
Joseph MacInnes (VML)


Neuroprosthetics

The Icelandic Vision Lab is working toward identifying potential benefits of neuroprosthetics involving intent control and sensory feedback, as well as their effects on the user. Further, the project aims to assess the relevance of current evaluation methods of lower-limb prostheses in the light of these new developments within the field and potentially develop new testing strategies to demonstrate their clinical benefits. The project is done in collaboration with Össur (supervision at Össur: Ásgeir Alexandersson, MD), the second largest prosthetics manufacturer in the world and a leader in research and development of advanced prosthetics.

People mainly involved in this project:

Árni Kristjánsson
Vigdís Vala Valgeirsdóttir


Synesthesia – Learning, Memory & Cognition

Synesthesia is a non-clinical condition in which people’s experiences of environmental stimuli elicit an automatic and consistent subjective experiences. The most common types of synesthesia are when a presentation of weekdays or months, letters, digits, or musical notes elicit a vivid experience of color, in the absence of any physical color stimulation. The experiences are often described as a sort of augmented reality, where normal perceptions are enriched with non-physical experiences of color, sounds, tastes or smells. With advances in cognitive neuroscience, synesthesia has captured the imagination of scientist interested in understanding the neuro-cognitive mechanisms underlying the peculiar phenomenon, and how this might affect behavior. The Icelandic Vision Lab has partnered with Dr. Thomas Alrik Sørensen, PI at the Centre for Cognitive Neuroscience, Aalborg University, to investigate how synesthetic associations are formed in brain, how they affect attention and memory, and how they are similar – as well as distinct from – visual and conceptual expertise. The research is supported by a grant from The Independent Research Fund Denmarkawarded to Dr. Sørensen. Readers interested in synesthesia, and those who may experience synesthesia can visit our website where they can fill out our survey, contact researchers for participation in the research or ask questions about the subject matter.

Árni was interviewed on Icelandic Public Radio, Rás 1, about synesthesia and related research. Listen here (Icelandic; 16th April, 2018).

People mainly involved in this project:

Árni Gunnar Ásgeirsson


Visual Factors in Dyslexia

Mounting evidence from our lab suggests that people with dyslexia have very specific problems with tasks that are thought to rely on the workings of high-level regions of the visual system. In a series of experiments, we are doing further studies on the role of visual factors in dyslexia, such as the role of fundamental visual properties, the type of visual processing, the role of experience, and the role of the number and complexity of visual objects. We are also putting the hypothesized deficient neural processing of high-level visual mechanisms in dyslexia to a direct test through the use of electroencephalography (EEG). Dr. Randi Starrfelt (University of Copenhagen, Denmark) is a member of a supervisory doctoral committee for this project.

People mainly involved in this project:

Hélène Devillez
Dýrleif Friðriksdóttir
Bahareh Jozranjbar
Árni Kristjánsson
Inga María Ólafsdóttir
Heida Maria Sigurdardottir


Visual Imagery

“I used to be a hyperphant and had something like HSAM [Highly Superior Autobiographical Memory]. I had a stroke in 2017 which left me with memory impairment (including SDAM [Severely Deficient Autobiographical Memory]) and aphantasia. I noticed immediately that my memories were gone and I knew something else was wrong, but it took a while to figure out that I’d lost my mind’s eye. It’s been devastating to me. I’ve lost my job and career, my sense of identity.”
– Anonymous, shared with permission

The person above has first-hand experience with two extremes of human imagination. Those with hyperphantasia have highly detailed and nearly photo-realistic visual imagery while those with aphantasia are unable to visualize at all – they have a blind mind‘s eye. We are interested in these extremes of human imagination and their relation to other cognitive abilities such as visual perception and visual memory.

People mainly involved in this project:

Árni Gunnar Ásgeirsson
Heida Maria Sigurdardottir
Sandra Thorudottir


Visual Statistics

Some of our research involves the representations of statistical variation in artificial displays and natural scenes. Recent evidence from our lab shows that observers can have a remarkably detailed representation of the distributions of stimuli in the environment. This work combines the disciplines of summary statistics, visual attention and sequential dependencies and priming in vision in addressing how we represent featural variation in the visual environment.  The project involves behavioral studies, computational modeling, and neurophysiological studies. This project is done in collaboration with Dr. Andrey Chetverikov (Donders Institute for Brain, Cognition and Behaviour, Radboud University, Netherlands) and Dr. Gianluca Campana (University of Padova, Italy).

People mainly involved in this project:

Sabrina Hansmann-Roth
Árni Kristjánsson
Mohsen Rafiei