|
Study of Medical Text Reading and Comprehension Through Eye-Tracking Fixations| title | Study of Medical Text Reading and Comprehension Through Eye-Tracking Fixations |
|---|
| start_date | 2024/11/15 |
|---|
| schedule | 14h-15h |
|---|
| online | no |
|---|
| location_info | A118 & en visio |
|---|
| summary | This study explores the reading and comprehension of medical texts by analyzing eye-tracking data, focusing on passages that present reading difficulties. Given the complex language often found in medical documents, simplifying these texts is crucial for improving patient understanding and engagement. Our objectives are twofold: first, to identify text passages and words that pose challenges during reading, and second, to train a large language model (LLM) to predict eye-tracking features based on textual input.
We conducted an eye-tracking experiment with participants from different backgrounds, analysing their reading behaviour on different text types, including original and simplified medical documents. Our results reveal significant differences in fixation duration, number of fixations and regression patterns between text types and participant groups, with specific eye-tracking features highlighting difficult words and passages.
Initial results from language model training show promising predictions for fixation measures, indicating that the model can help identify text complexity. The next steps will be to extend this framework to a multitasking approach, in which the model simultaneously predicts eye-tracking characteristics and classifies text and reader types, paving the way for a more nuanced understanding of readability in medical contexts. |
|---|
| responsibles | Gao, Cui |
|---|
Workflow history| from state (1) | to state | comment | date |
|---|
| submitted | published | | 2024/11/07 08:51 UTC |
| |
|