|
During reading the eyes do not smoothly slide trough the sentence, they rather show an alternating pattern of stable phases (so called fixations) and jumps (so called saccades). Whereas most of the saccades are in the direction of reading (i.e., from left to right in most Western European writing systems), some of the saccades go in the opposite direction, moving the eyes back to sentence material that has been fixated earlier. These so called regressive saccades occur frequently in the context of difficulties in language processing but their exact functional role in reading is largely unclear. Especially, it is still an open question in which cases the eyes just show increased fixation durations and in which cases the eyes actually regress. The present thesis focuses on the functional role of regressive eye movements during reading. In the first part, the limitations of the existing literature in the field of regressive eye movements are discussed. After that, a new account is presented which is called the Information Gathering Framework. This framework postulates at its core that each regressive eye movement has a function, namely to gather additional information about a word that has been processed earlier. Based on this hypothesis a detailed architecture of the model is outlined and it is shown how the model may account for empirical findings from reading research that have been reported in the literature. Also further predictions of the model are discussed. In the second part of the thesis, 2 experiments are presented that were developed to explicitly test the predictions of the Information Gathering Model. The first experiment focuses on the influence of different tasks on regression behavior using eyetracking. The second experiment (consisting of two sub experiments) focuses on the neural correlates of regressive eye movements using combined fMRI / eyetracking measurements. The results of both experiments provide clear empirical evidence in favor of the assumptions of the presented model. In particular, ...
|
