Optical Illusions and Human Visual System: Can we reveal more? Imaging Science Innovative Student Micro-Grant Proposal 2011

Transcription

1 Optical Illusions and Human Visual System: Can we reveal more? Imaging Science Innovative Student Micro-Grant Proposal 2011 Prepared By: Principal Investigator: Siddharth Khullar 1,4, Ph.D. Candidate Faculty Mentor: Dr. Nathan Cahill 1,2, Professor Co-Investigators: Dr. Jeff Pelz 1, Professor Dr. Daniel Phillips 3, Professor External Consultants: Dr. Vince Calhoun 4 Dr. Andrew Michael 4 1. Chester F. Carlson Center for Imaging Science 54 Lomb Memorial Drive 2. Department of Statistics and Mathematics 79 Lomb Memorial Dr 3. Center for Applied and Computational Mathematics 85 Lomb Memorial Dr 4. Mind Research Network for Neurodiagnostic Discovery and Mental Illness 1101 Yale Blvd. NE Albuquerque, NM /3/2011 1

2 1. Abstract The act of perceiving objects seems quite trivial, making it easy for us to turn a blind eye to the sophisticated and poorly understood machinery behind this process. Illusions may be described as the stimuli that exist at the extremes of our visual system s evolution profile, making them difficult to handle. There is no unique cause for optical illusions in the visual system; they may be a result of assumptions made by the visual system or interestingly, represent an active recalibration. The primary purpose of this research would be to provide a comprehensive insight about the response of the visual system (eye to-cortex) to various types of optical illusions. Preliminary data would be collected from a group of healthy subjects with normal (or corrected-to-normal) vision while they are presented with a mixture of illusions interleaved with control stimuli. The data would comprise of neural activity recordings using a 32-channel EEG system in addition to eyetracking data, primarily spontaneous saccades. Proposed here is a hybrid framework to find neural activity and identify brain regions responsible for perception of these illusions. We expect to find resulting patterns in the data that are uniquely associated with different types of illusory motions such as the peripheral drift rotating snakes and the aftereffect motion waterfall effect. 2. Dollar Request & Funding Dates Total - $6000 Desired Funding Dates: March 2011 June Proposal 3.1 Scientific Justification Interestingly, an optical illusion can be defined in a general sense. For example, as humans we are completely unaware of the edges present in the visual field, that is, our angle of vision is limited, making this lack of information eligible to be known as an illusion [1]. Another good example that appears in our everyday lives is that of halftone patterns perceived as continuous tones due to spatial vision capabilities of the eye. Intuitively, it is hard to define a single perfect example that describes visual illusions, rather numerous categories are preferred. The purpose of this research is partially inspired by a recent fmri (functional magnetic resonance imaging) study by Kuriki et al. [2]. In the past two decades, there has been increased interest in understanding the neural basis of illusory perception evoked from static images, and unfortunately, until this recent study by Kuriki et al. [2], there was no evidence showing what region of the Visual Cortex is responsible for this illusory motion. They examined changes in neural activity in the motion sensitive areas of the human visual cortex, that is, hmt+ when a static illusory-motion image, commonly known as Rotating Snakes was presented to subjects. For comparison purposes, the neural 2

3 activity due the test stimulus (Fig. 1(a)) was compared against a control stimulus (Fig. 1(b)) that produced no illusion. Figure 1: Illustration of the Test Stimuli and Control Stimuli used in the experiment by Kuriki et al [2]. The test stimulus shown in Fig. 1(a) consists of snakes with two different orders: (1) BLUE-WHITE-YELLOW-BLACK (appears to rotate clockwise) and (2) BLUE- BLACK-YELLOW-WHITE appears to rotate anti-clockwise. It was earlier stated that the orientation of rotation varied from person to person, but it seems the rotation depends on the direction in which the gray values decrease/increase, if the grey values are coupled as black-to-dark gray and white-to-light gray, the illusion appears clockwise whereas if they are coupled as white-to-dark gray and black-to-light gray, the illusion appears anticlockwise. If they are coupled in any other order, the illusion disappears as seen in Fig. 1(b). We think it would be more rewarding to understand these illusions at a much higher temporal resolution by using EEG recordings from different parts of the brain and eyetracking equipment in contrast to Kuriki s method [2] that uses fmri recordings. With EEG data, it would be possible to strongly link different areas in the brain that function concurrently or collinearly to produce the illusory effect. The EEG activity can then be linked with eye-movement patterns (random or guided) that could reveal important information regarding related links between low-level and high-level vision. 3.2 Proposed Methodology The Experiment Different test stimuli, as shown in Figure 2, have been selected for this experiment that cause different illusions: (1) waterfall aftereffect illusion and (2) peripheral drift rotating snake illusion. Before starting the stimulus presentation, the rest screen shall be presented for 12 seconds. The presentation of snakes stimulus for 12 seconds should be followed by a presentation of a uniform gray screen presentation (rest). Whereas, the waterfall (task) stimulus is presented for 12 seconds followed by appearance of an image of a textured object (a rock) in order to perceive the after effect. The waterfallstimulus pair is then followed by rest screen (gray) in order to let the BOLD response return to a stable unexcited state. The final order of the stimulus pairs is : snakes, rest, 3

4 waterfall, rock, rest, snakes, rest, and so forth. This sequence is repeated 3 times in each run as shown in Figure 2. Figure 2: Proposed block design involving alternating stimuli that induce different types of illusory motions. The striped waterfall illusion rotates for 12 seconds followed by image of the rock, where the illusion is perceived. Image Sources: (1) Lecture Slides by Dr. Jeff Pelz (Human Vision) (2) Rock: sugarmtnfarm.com/blog/2007/06/round-rock.html Analysis Methods EEG and eye-tracking data will be recorded while the subjects perform the aforementioned experiment. We choose to apply robust analysis methodologies such as ICA (Independent Component Analysis) [3]. ICA is a source separation method commonly applied to neuroimaging studies for estimating the spatial source of different functionally connected regions. For analysis of the EEG data, we choose ICA for analysis given its ability to separate different components associated with independent neural activity in different regions of the brain. This will help segregate areas responsible for processing different types of illusions in addition to the correlations between them. As it is initially hypothesized that the two stimuli might evoke responses in different regions of the brain, it would be interesting to see how the ICA separates these responses. If there are some regions in the brain that show significant activation during both the tasks, ICA will help reveal new and useful facts about the neural correlates between different illusions in addition to functional networks in the brain associated with a specific type of illusion. As an extension to this work, it may be possible to use the obtained results to create a model that is complementary to the process of perceiving illusions. In other words, it may be possible to create or utilize a specific type of visual stimuli/process to help counter-act the illusory motion perceived due to these illusions. As a long term advantage, this application may be useful in alleviating vertigo that is a common medical problem experienced by millions of people around the world. References 1. Eagleman D.M., Visual illusions and neurobiology, Nat. Rev. Neuroscience. Vol. 2, pp (2001). 2. Kuriki I. and Ashida H., Functional brain imaging of Rotating Snakes illusion by fmri, J. Vision, 8(10):16, pp. 1-10, (2008). 3. Eichele T, Calhoun VD, Debener S, Mining EEG-fMRI using ICA, Int. J. Psychophysiol.; 73(1):53-61 (2009). 4

5 4. Budget Request A preliminary budget is presented below. The actual budget may differ slightly depending on the requirements during the course of the project. S No. Description Amount (in USD) 1 Stipend for 2 undergraduate students (10-15 hrs/week) $ 2000 X 2 2 Conference travel to present research outcomes $ Subject/volunteer benefits (hourly pay) for ~ 20 subjects $ Equipment (External HDD etc.) $ 200 Total $ Project Deliverables Anticipated timeline: March 2011 to May 2011 (Spring ) The work done in the course of the project will be made available to RIT and CIS students in form of a technical report or a white paper. This research will be presented at various group meetings such as the CIS weekly seminar, weekly IS&T meetings and the Imagine RIT innovation festival. This will help gain interest and support for writing grants to external funding agencies such as NSF and NIH for further extension of research. We plan to achieve the following milestones towards the end of this research project: 1. A strong understanding of optical illusions in conjunction with EEG recordings and eye-movement patterns for a group of subjects. 2. Identifying unique patterns that may be strongly associated with perceiving illusory motion across a range of subjects and patterns that are unique to the whole group. 3. Write a full-scale conference paper highlighting the findings of this research and submit to a prestigious signal processing/computer vision conference such as CVPR 2012 or ICASSP Make the recorded EEG and eye-tracking data available openly for free use by researchers in the field of imaging and signal processing. 5. Utilize the visual experiment and collect fmri data using a different set of subjects and report spatial localizations (and connections) within the brain responsible for processing illusory motion. 5