iTrack: instrumented mobile electrooculography (EOG) eye-tracking in older adults and Parkinson’s disease

Stuart, Samuel, Hickey, Aodhán, Galna, Brook, Lord, Sue, Rochester, Lynn and Godfrey, Alan (2017) iTrack: instrumented mobile electrooculography (EOG) eye-tracking in older adults and Parkinson’s disease. Physiological Measurement, 38 (1). N16-N31. ISSN 0967-3334

[img]
Preview
Text
iTrack.pdf - Accepted Version

Download (926kB) | Preview
Official URL: https://doi.org/10.1088/1361-6579/38/1/N16

Abstract

Detection of saccades (fast eye-movements) within raw mobile electrooculography (EOG) data involves complex algorithms which typically process data acquired during seated static tasks only. Processing of data during dynamic tasks such as walking is relatively rare and complex, particularly in older adults or people with Parkinson's disease (PD). Development of algorithms that can be easily implemented to detect saccades is required. This study aimed to develop an algorithm for the detection and measurement of saccades in EOG data during static (sitting) and dynamic (walking) tasks, in older adults and PD. Eye-tracking via mobile EOG and infra-red (IR) eye-tracker (with video) was performed with a group of older adults (n  =  10) and PD participants (n  =  10) (≥50 years). Horizontal saccades made between targets set 5°, 10° and 15° apart were first measured while seated. Horizontal saccades were then measured while a participant walked and executed a 40° turn left and right. The EOG algorithm was evaluated by comparing the number of correct saccade detections and agreement (ICC2,1) between output from visual inspection of eye-tracker videos and IR eye-tracker. The EOG algorithm detected 75–92% of saccades compared to video inspection and IR output during static testing, with fair to excellent agreement (ICC2,1 0.49–0.93). However, during walking EOG saccade detection reduced to 42–88% compared to video inspection or IR output, with poor to excellent (ICC2,1 0.13–0.88) agreement between methodologies. The algorithm was robust during seated testing but less so during walking, which was likely due to increased measurement and analysis error with a dynamic task. Future studies may consider a combination of EOG and IR for comprehensive measurement.

Item Type: Article
Subjects: B900 Others in Subjects allied to Medicine
G900 Others in Mathematical and Computing Sciences
Department: Faculties > Engineering and Environment > Computer and Information Sciences
Depositing User: Becky Skoyles
Date Deposited: 20 Apr 2018 12:15
Last Modified: 01 Aug 2021 08:47
URI: http://nrl.northumbria.ac.uk/id/eprint/34047

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

View more statistics