Thesis title:
Discourse Analysis and Eye-Tracking for Detecting Cancer-Related Cognitive Impairment.
Abstract:
Some cancers are becoming long-term diseases and life after cancer is now a public health issue. Breast cancer is the most common cancer in women and is also the best treated.
Cancers and their treatments can cause subtle cognitive disorders called Cancer-Related Cognitive Impairment (CRCI). These disorders affect up to 75% of patients and can last years following completion of treatments for cancer (chemotherapy, radiotherapy, and surgery). It negatively impacts patients’ quality of life, hindering their return to work or limiting their participation in social activities. Patients complain of distractibility, forgetfulness, difficulty with double-tasking, and difficulties with language and communication (word-retrieval and syntactic planning difficulties).
Although objective tests sometimes reveal an impairment of attention, memory, executive functions and language, there is a discrepancy between the intensity of the complaint and the neuropsychological tests scores. This discrepancy could be explained by a lack of sensitivity and ecological validity of the tests, designed for more severe disorders. Nowadays, the cognitive mechanisms implied in the CRCI (attention, memory, or executive functions) remain poorly identified. Moreover, the lack of specific diagnostic tools prevents some patients from accessing appropriate support and slows down the progress of our understanding of these disorders.
One of the most common manifestations of CRCI, the word-retrieval difficulties (or anomia), is particularly affected by the lack of sensitivity and ecology of the tests, as it is assessed by fluency and naming tasks. In speech, anomia is characterized by interruptions in the flow of speech (pauses, repetitions, or rephrasing of a sentence) during the search for a word. The analysis of these speech accidents, called disfluencies, can not only detect a fine modification of the language, but can also contribute to the identification of the underlying cognitive mechanisms.
Speech and gaze are closely linked, and many studies have shown the interest of eye-tracking technologies to measure certain cognitive functions. By combining discourse analysis and eye-tracking, we aim to overcome the lack of tools for the assessment of the CRCI. Our objectives are (i) to identify the linguistic difficulties using discourse analysis, (ii) to better understand the underlying cognitive mechanisms, (iii) to develop a diagnostic tool based on the analysis of disfluencies to provide appropriate care.
Thesis supervisors:
Dr Karen T. REILLY (Lyon Neuroscience Research Centre UMR 5292), Pr Sophie JACQUIN-COURTOIS, M.D. (Lyon Neuroscience Research Centre UMR 5292 ; Hospices Civils de Lyon) & Pr Fabrice HIRSCH (Université Paul-Valéry, Praxiling UMR 5267 ).
