Analysis for Extracted Features of Pupil Light Reflex to Chromatic Stimuli in Alzheimer’s Patients

• Main Authors: Wioletta Nowak, Minoru Nakayama, Tomasz Kręcicki, Elzbieta Trypka, Artur Andrzejak, Andrzej Hachoł
• Format: Article
• Language: English
• Published: European Alliance for Innovation (EAI) 2019-02-01
• Series: EAI Endorsed Transactions on Pervasive Health and Technology
• Subjects: pupil light reflex; alzheimer’s disease; light colour; feature extraction; classification
• Online Access: https://eudl.eu/pdf/10.4108/eai.13-7-2018.161750
• ISSN: 2411-7145

INTRODUCTION: Some Alzheimer’s Disease (AD) patients respond to chromatic light stimulus, which mayinfluence intrinsically photosensitive retinal ganglion cells (ipRGCs), due to factors common to both AD andAge-Related Macular Disease (AMD).OBJECTIVES: In this study, short light pulses of three colours were introduced to a novel diagnostic procesure for AD patients such as classification techniques using waveform features of pupil light reflexs (PLRs), and their prediction performances were evaluated.METHOD: PLRs to 1s pulses of red, blue and white stimuli shown at high and low photopic levels followedby a 7s restoration process were recorded after the stimuli were shown to 7 AD patients and 12 non-ADparticipants (aged 42-89). Features of waveform shapes of PLRs in 5 dimensions and 15 features of PLRs were extracted.RESULTS: In a classification analysis, most non-AD participants were correctly identified using the samelevel of performance we reported when PLRs for red and blue stimuli were used to measure the performanceof AD patients. There were significant differences in some of the features of PLRs extracted from the twogroups (AD and non-AD participants), particularly with the features for blue light stimuli in high brightness,which produced significant reactions in AD patients. The classification performance of using 15 features ofthe response to blue light stimuli was the highest among responses for all three colours, and was higher than the performance using the procedure in the previous study. Also, a few of the features extracted using the three colours of stimuli changed significantly across age ranges (70 and under, 71-80, and over 80), so these may indicate factors related to ageing.CONCLUSION: These results confirm that some specific features of PLRs, in particular the response to blue light, can indicate the existence of AD in patients. Also, a few of the features may reveal factors related toageing during evaluations which use PLRs. This evidence may help in the better understanding of features of PLRs.