Pupillometry can be used as a reliable physiological measure of cognitive processing [15], [19], [36], [39]. This assumption has been supported by a body of research demonstrating how pupil size fluctuates in response to characteristics of cognitive processing. This line of research is instigated by the pioneering studies of Hess and Polt [26], [27] who reported how pupil size increases when participants solve multiplication equations of increasing difficulty. Based on these studies, Kahneman and Beatty [33] have reported that pupil size increases in response to increased difficulty of a working memory task consisting of repeating a series of numbers, either in a forward or reverse order. These findings have been replicated by a number of studies demonstrating how pupil size increases as cognitive load increases [2], [9], [23], [43], [55]. The cognitive load hypothesis has been also supported by research in linguistics demonstrating that pupil size increases with difficulties in processing ambiguous meanings of sentences [29] and difficulties in reading sentences that differ in syntactic complexity [32]. Research has also reported how pupil size is sensitive to the load of lexical translation during simultaneous interpretation [5]. Taken together, cognitive processing can be assessed with pupillometry. In this study, we thus investigated whether pupillometry may assess verbal fluency in older adults.

To the best of our knowledge, prior research has not assessed verbal fluency using pupillometry in older adults. That being said, some research has used pupillometry to assess other cognitive domains in older adults. For instance, Piquado et al. [44] have assessed whether pupillary size increases with cognitive load in older adults on a task requiring the retention of digits increasing in length. Results demonstrated increased pupil size with the increased length of digits. The effect of cognitive load on pupil size in aging was also assessed by Allard et al. [1] who presented older adults with facial stimuli of varying emotions. The authors also assessed the mood of participants during viewing the stimuli. Results demonstrated larger pupil size during viewing faces in participants in neutral mood compared to participants in positive or negative mood. The authors attributed this variation in pupil size to cognitive regulatory strategies deployed by older adults when faced with stressful emotional contexts (e.g., when exposed to faces depicting anger). The paucity of research using pupillometry to assess cognitive processing in older adults can be attributed to several factors such as a decrease of pupil size and a restricted range in pupil reaction in aging [44]. Furthermore, older adults tend to demonstrate low pupil sensitivity when cognitive load slightly changes [54]. These factors may explain why fewer research studies have used pupillometry to assess cognitive processing in general and, more specifically verbal fluency, in older adults.

Prior studies have, however, used pupillometry to assess verbal fluency in younger adults. More specifically, El Haj et al. [18] measured pupil dilation during a verbal fluency task and during a control task in young participants (Mean age = 23.55 years). On the verbal fluency task, participants were invited to generate during a one-minute period as many words as possible beginning with the letter “P”. On a control task, participants were invited to count aloud during a one-minute duration. In both tasks, pupil activity was recorded with eye-tracking glasses. Results demonstrated larger pupil size during the verbal fluency condition compared with the counting condition. The larger pupil size during the verbal fluency condition can be attributed to the cognitive load of the verbal fluency task, which involves both linguistic processing and executive function (e.g., inhibiting inappropriate responses). Based on El Haj et al. [18] study, we assess in the present study whether verbal fluency would also influence pupil size changes in older adults. This issue is important not only because there is a lack of linguistic research in older adults using pupillometry but also because there is a need to provide reliable physiological assessment of verbal fluency in older adults in neuropsychological assessment. Verbal fluency tasks are typically assessed using paper and pencil in neuropsychological testing of older adults and there is a need to provide a reliable behavioral index of verbal fluency, and linguistic processing in general, in neuropsychological assessment of older adults. Pupillometry is an ecological and noninvasive physiological tool that may be used as an assessment of verbal processing in older adults.

To summarize, little, if any, attempt has been made to assess verbal fluency with pupillometry in older adults. We thus evaluated pupil size during a verbal fluency and a control task in older adults. More specifically, we recorded pupil size during a verbal fluency task in which we invited participants to retrieve as many words as possible beginning with the letter “P”. Note that we measured verbal fluency only regarding the letter “P” because this letter is mainly assessed in phonemic fluency tasks in French populations [3]. As a control condition, we invited participants to count aloud. We considered this control condition because, like verbal fluency, counting aloud involves verbal behavior, so any potential differences in pupil dilation between the two conditions would not be attributed to verbal behavior. We expected a larger pupil size during the verbal fluency task than the control task, as verbal fluency requires more cognitive load (e.g., inhibiting production of inappropriate words) than counting. To investigate executive function, we assessed whether pupil size would be correlated with inhibitory failures during the verbal fluency task. These failures were assessed in terms of perseverative errors (i.e., the repetition of previously-produced words during the verbal fluency task).