Beyond eye gaze: What else can eyetracking reveal about cognition and cognitive development?

Abstract

This review provides an introduction to two eyetracking measures that can be used to study cognitive development and plasticity: pupil dilation and spontaneous blink rate. We begin by outlining the rich history of gaze analysis, which can reveal the current focus of attention as well as cognitive strategies. We then turn to the two lesser-utilized ocular measures. Pupil dilation is modulated by the brain's locus coeruleus-norepinephrine system, which controls physiological arousal and attention, and has been used as a measure of subjective task difficulty, mental effort, and neural gain. Spontaneous eyeblink rate correlates with levels of dopamine in the central nervous system, and can reveal processes underlying learning and goal-directed behavior. Taken together, gaze, pupil dilation, and blink rate are three non-invasive and complementary measures of cognition with high temporal resolution and well-understood neural foundations. Here we review the neural foundations of pupil dilation and blink rate, provide examples of their usage, describe analytic methods and methodological considerations, and discuss their potential for research on learning, cognitive development, and plasticity.

Keywords: Blink rate; Children; Eyetracking; Pupil dilation; Pupillometry; Saccades.

Fig. 1

Eye muscles responsible for eye movements and pupil dilation and contraction. a) Superior view of the eye. The superior and inferior rectus muscles are responsible for the eye’s vertical movements, whereas the lateral and medial rectus muscles control horizontal movements. Adapted with permission from Eds. Levin et al., (2011). b) Top: The dilator pupillae muscle dilates the pupil and is controlled by sympathetic fibers. Bottom: The sphincter pupillae muscle contracts the pupil and is controlled by parasympathetic fibers. The balance between the activation of the dilator and sphincter pupillae muscles dictates pupil diameters.

Fig. 2

Gaze analyses in developmental research on attentional capture in infancy and visuospatial reasoning in children. a) While viewing speakers’ faces, four-month olds spent a greater proportion-of-total-looking-time (PTLT) on a speaker's eyes, whereas 8- to 12-month olds spent greater PTLT on a speaker's mouth. In adulthood, the balance shifts back to a speaker's eyes. Reprinted with permission from Lewkowicz and Hansen-Tift (2012). b) In a developmental comparison of matrix reasoning, the authors defined an “encoding” sequence as three subsequent fixations along a row or column of the matrix problem space. Using a median split by performance, higher-performing 5–6 year-olds demonstrated approximately the same encoding prevalence as 7–8 year-olds (left). Similarly, children who received feedback on how to complete the puzzles demonstrated more encoding behavior than those who did not (right). Reprinted with permission from Chen et al. (2016).

Fig. 3

Temporal coupling between pupil diameter and firing of a single LC neuron of a monkey during performance of a signal-detection task. The relationship between LC firing and pupil diameter is mediated through the projection of the LC to the Edinger-Westphal nucleus, the origin of the pupil’s parasympathetic constricting fibers, and through the influence of the LC-NE system on sympathetic nervous activity, which promotes pupil dilation. Reprinted with permission from Rajkowski et al. (1993).

Fig. 4

Anatomy of the autonomic nervous system and its sympathetic (fight-or-flight) and parasympathetic (rest-and-digest) branches. Post-ganglionic activity is mostly mediated by NE in the sympathetic branch and by acetylcholine in the parasympathetic branch. Many organs receive inputs from the sympathetic and parasympathetic branches, in which case functions are often reciprocal, as with pupil dilation (sympathetic dilates, parasympathetic constricts) or heart rate (sympathetic accelerates, parasympathetic slows). Reprinted with permission from the Merck Manual Professional Version, known as the Merck Manual in the US and Canada and the MSD Manual in the rest of the world, edited by Robert Porter. Copyright 2016 by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co, Inc, Kenilworth, NJ. Available at http://www.merckmanuals.com/professional. Accessed April 27, 2016.

Fig. 5

NE and DA pathways in the brain and their relationship to cognitive performance. a) The LC is the only source of cortical NE but has widespread and highly specific connections throughout the entire nervous system. The LC-NE system promotes physiological arousal and is crucial for a variety of cognitive functions, such as attention, memory, and decision making. b) DA cells in ventral tegmental area (VTA) innervate the mesocorticolimbic pathway that projects to limbic and cortical regions. In the mesostriatal pathway, the striatum receives input from DA cells of the substantia nigra. Reprinted with permission from Breedlove et al. (2010). c) Task performance is optimal at intermediate levels of NE, at which task-relevant stimuli elicit pronounced phasic LC responses. Low levels of NE are associated with inattentive behavior and drowsiness, and high levels with distractibility. Adapted with permission from Aston-Jones et al. (1999). d) Just as for NE, the relationship between DA levels and cognitive control performance can be described by a quadratic function. Specifically, this inverted U-shape relationship has been widely documented for D1 receptor activity and working memory performance. Adapted with permission from Goldman-Rakic et al. (2000).

Fig. 6

Pupil dilation scales with task difficulty in a variety of cognitive domains. a) Short-term memory: Digit-span task. Subjects saw 3–8 digits, presented sequentially for one second each, and attempted to recall all digits after a retention interval of 3 s. Pupil dilation increased as a function of short-term memory load. Reprinted with permission from Klingner et al. (2011). b) Working memory: Multiplication task. Subjects were asked to mentally multiply two visually presented numbers. The numbers were smallest in the “easy” condition, bigger in the “medium” condition, and biggest in the “hard” condition. Pupil dilation scaled with task difficulty and remained elevated for several seconds after stimulus presentation. Reprinted with permission from Klingner (2010). c) Task-relevant processing: Oddball task. Subjects listened to a stream of auditory stimuli and were instructed to press a button in response to target tones only. Target tones (1500 Hz) made up 10% of the presented stimuli, 80% were standard stimuli (1000 Hz), and 10% were novel stimuli (bells, whistles, horns, etc.). There was no sign of pupil dilation in response to standard stimuli. Novel stimuli elicited a pronounced pupil dilation of more than 0.5 millimeters, but target tones elicited a much larger response of 2 millimeters, reflecting selective orientation toward task-relevant stimuli. Reprinted with permission from Book et al. (2008). d) Cognitive control: Stroop task. Subjects were asked to name the color of 320 letter combinations presented for 2 s. In congruent trials, the colored letters formed the name of the color, whereas in incongruent trials, the letters formed the name of another color. Non-color words were used as a control condition. Pupil dilation was reduced in congruent trials relative to non-color words and was increased in incongruent trials, suggesting that pupil dilation is a sensitive measure of cognitive control. Reprinted with permission from Laeng et al. (2010).

Fig. 7

Examples for the use of pupillometry in development. a) Object permanence. Ten-month-old infants saw drawbridges that, by rotating, occluded a box behind. Infants’ pupils responded to the rotation of the drawbridge, revealed by a main effect of rotation (180° or 120°), and to the presence of a box, revealed by a main effect of the presence or absence of the box, but did not respond to the violation of the principles of object permanence (box present and 180°), as would be revealed by an interaction between both. Adapted with permission from Sirois and Jackson (2012). b) Short-term memory. Differences in pupil dilation between children and adults while listening to long sequences of to-be-recalled digits. The premature drop in children’s pupil diameters before the end of the sequence suggests that their worse recall performance might be caused by a lack of attention during encoding and a failure to allocate sufficient cognitive resources. Reprinted with permission from Johnson et al. (2014).

Fig. 8

Relationship between blink rate and DA receptor activity. a) Blink rate is positively related with PET measures of D₂-like receptor availability in the ventral striatum (white circles) and caudate nucleus (black circles), but not putamen (gray circles). These relationships were not observed with D₁-like receptors (not shown). b) Statistical map (p-values) of the voxelwise linear regression of blink rate on D₂-like receptor availability from (a) overlaid on the striatal volume of the vervet monkey's MRI template. Adapted with permission from Groman et al. (2014). c) Systemic administration of apomorphine, a non-selective DA agonist, increased blink rate in a dose-dependent manner (orange lines) above baseline levels (saline administration, black line) in marmosets. d) This effect was only reversed with the administration of SCH39166, a D₁-antagonist (blue line), but not with the administration of haloperidol, a D₂-antagonist (not shown). Adapted with permission from Kotani et al. (2016).

Fig. 9

Measurement of spontaneous eyeblink rate in developmental studies. a) Blink rate during performance of a task that requires rule switching flexibility can be used to differentiate between typically developing children and children with Tourette syndrome (TS), a condition associated with elevated levels of DA. Patients showed higher average blink rate at rest as did the control subjects, and also did not show task difficulty-related increases in blink rate. Adapted with permission from Tharp et al. (2015). b) Relationship between blink rate at rest and age-related differences in reward-seeking behavior on a risky decision-making task. Higher blink rate predicted increased use of a gain-maximizing strategy for adolescents but not for adults. Adapted with permission from Barkley-Levenson and Galván (2016).