Concussion Detection Research Tool Codi-Lee Hayes Samantha Mearns Rebecca Yaffe Dr. Thirimacho Bourlai Dr. Aaron Monseau
Problem Statement Concussions if go undetected could lead to more serious brain injury, especially if the sufferer does not limit their physical activity after an incident. Brain imaging technology cannot detect concussions because there are no signs of an apparent injury. Concussions are usually diagnosed through a comprehensive eye exam by a trained doctor or can be diagnosed through the use of a side line test. A system or device that could track the patient’s eyes and compare their movements against a trend line of normal eye movements to automatically determine the likelihood of a patient having a concussion would solve this problem.
Introduction Find a pattern between individuals with a concussion and a nystagmus. Find a pattern between individuals with a concussion and a nystagmus. A nystagmus is an involuntary, repetitive eye movement due to irregular patterns in the brain that control eye movement. A nystagmus is an involuntary, repetitive eye movement due to irregular patterns in the brain that control eye movement. A controlled collection was set up in order to record the eye movements of 10 individuals undergoing a nystagmus test using a Canon 5D Mark iii. A controlled collection was set up in order to record the eye movements of 10 individuals undergoing a nystagmus test using a Canon 5D Mark iii. A GUI was created using the software MATLAB in order to load, separate, track and collect the pupil patterns from the video footage. A GUI was created using the software MATLAB in order to load, separate, track and collect the pupil patterns from the video footage. This data was recorded on a graph in order to map the pixel distances of the pupils throughout the video This data was recorded on a graph in order to map the pixel distances of the pupils throughout the video These pixel values established a trend line of normal eye movements that will be compared to abnormal eye movements in order to diagnose a concussion These pixel values established a trend line of normal eye movements that will be compared to abnormal eye movements in order to diagnose a concussion
Video of patient with a nystagmus
Goals Determine a difference between normal eye movements and the eye movements of patients suffering from recent brain trauma (concussion) Create a system that maps eye movements for concussion detection research Gain funding for future development for the final product
Objectives Perform a collection of eye movements from 10 subjects Perform a collection of eye movements from 10 subjects Create a graphical user interface (GUI) to obtain data Create a graphical representation of the eye movements Find a trend line of normal eye movements based on our data
Collection
GUI The software MATLAB R2013A was used The GUI was implemented in order to create a concussion detection research tool and consists of 4 main functions. Allows the user to load the video footage frame-by-frame, trace the center of the pupils manually from the left to the right pupil, go to the next/previous frame and finally generate a plot based on the pixel distances that are saved into a text file.
Use Case
GUI Flow diagram
Data Flow Diagram
Load Video Function This function loads the video frame by frame into the GUI in order to start collecting pupil data. Saves the traced pixel values of the left and right pupil and also features error detection. There is an error detection message that will be displayed if the user doesn’t correctly store the left pupil’s data first.
Next Function This function allows the user to go to the next frame of the video once the data for the left and right pupil has been traced. This allows the user to navigate frame by frame once the data has been collected for each pupil and contains error detection to make sure the data points are valid.
Previous Function This function allows the user to go to the previous frame and record values by overwriting the previous values. This function also contains error detection so that the user obtains the data from the left pupil before the right in order to make the research consistent. This function traces the frame the user is currently obtaining data in order to correctly override the previous data that was stored within the matrix.
Generate Plot Function This function correctly plots the pixel values for both the left and right pupils. These values are based off of the crosshair points selected by the user. Plot is based on pixel distances
GUI Demo
Accomplishments The Concussion Detection Research Tool we created successfully loads a video of eye movements and allows the user to collect the pixel distances of each eye for all frames in the video Our research shows an obvious difference in normal eye movements and the eye movements of a patient suffering from a concussion The tool is easy to use so that it can be used for future research once funding is available for further research and analysis.
Graphical Data Normal Eye Movements Abnormal Eye Movements
Business Potential Sports Community Accurately and quickly determine if an athlete needs to be removed from a game which could save them from further injury. Efficiently diagnose a concussion Medical Community Doctors, nurses, coaches, EMT workers, etc. will have the ability to diagnose the likelihood of a concussion quickly. This research could potentially save a hospitals time and money. Reduced waiting times and fast/accurate diagnosis at a low cost.
Future Development Make the code completely autonomous Package smaller (app) Use our tool to give data to actually make a final product
Questions?