Cortical modulation of pupillary function: systematic review

Background The pupillary light reflex is the main mechanism that regulates the pupillary diameter; it is controlled by the autonomic system and mediated by subcortical pathways. In addition, cognitive and emotional processes influence pupillary function due to input from cortical innervation, but the exact circuits remain poorly understood. We performed a systematic review to evaluate the mechanisms behind pupillary changes associated with cognitive efforts and processing of emotions and to investigate the cerebral areas involved in cortical modulation of the pupillary light reflex. Methodology We searched multiple databases until November 2018 for studies on cortical modulation of pupillary function in humans and non-human primates. Of 8,809 papers screened, 258 studies were included. Results Most investigators focused on pupillary dilatation and/or constriction as an index of cognitive and emotional processing, evaluating how changes in pupillary diameter reflect levels of attention and arousal. Only few tried to correlate specific cerebral areas to pupillary changes, using either cortical activation models (employing micro-stimulation of cortical structures in non-human primates) or cortical lesion models (e.g., investigating patients with stroke and damage to salient cortical and/or subcortical areas). Results suggest the involvement of several cortical regions, including the insular cortex (Brodmann areas 13 and 16), the frontal eye field (Brodmann area 8) and the prefrontal cortex (Brodmann areas 11 and 25), and of subcortical structures such as the locus coeruleus and the superior colliculus. Conclusions Pupillary dilatation occurs with many kinds of mental or emotional processes, following sympathetic activation or parasympathetic inhibition. Conversely, pupillary constriction may occur with anticipation of a bright stimulus (even in its absence) and relies on a parasympathetic activation. All these reactions are controlled by subcortical and cortical structures that are directly or indirectly connected to the brainstem pupillary innervation system.


INTRODUCTION
The pupillary light reflex is a polysynaptic reflex that requires cranial nerves II and III, as well as central brainstem connections (Kawasaki, 1999). Light falling into one eye stimulates retinal photoreceptors, bipolar cells and subsequently retinal ganglion cells whose axons form the optic nerve. Some of these axons terminate in the pretectum of the mesencephalon and pretectal neurons project further to the Edinger-Westphal nuclei. Then, preganglionic parasympathetic axons synapse with ciliary ganglion neurons which in turn send postganglionic axons to innervate the pupillary constrictor muscles of both eyes. Conversely, pupillary dilatation relies on the sympathetic system which consists of pre-ganglionic fibers projecting from the hypothalamus to the superior cervical ganglion and post-ganglionic fibers projecting to the iris dilator muscles, via ciliary nerves (Kawasaki, 1999).
In addition to brainstem pathways, there exists also a cortical component of pupillary innervation. For instance, emotional responses such as surprise and cognitive processes such as decision making, memory recall and mental arithmetic may produce pupillary dilation (Steinhauer, Condray & Kasparek, 2000;Simpson & Hale, 1969;De Gee, Knapen & Donner, 2014). Pupillary function may be assessed as changes in pupillary size relative to resting state diameter or alterations of the light reflex in terms of reflex amplitude and latency (i.e., time from light stimulus to pupillary constriction). Cognitive scientists and psychologists have used measurements of pupillary diameters since the 1960ies to monitor mental processes in healthy volunteers and people with a wide range of neurological and psychiatric disorders, including Alzheimer's disease, autism and anxiety (Bittner et al., 2014;Lim et al., 2016;Bakes, Bradshaw & Szabadi, 1990;Krach et al., 2015). Testing of emotional processes usually involves neutral versus emotionally salient stimuli, e.g., pictures of everyday life objects versus pictures evoking sadness, anger or happiness, whereas cognitive processes are investigated with tasks such as arithmetic calculations and memory recall tests (Steinhauer, Condray & Kasparek, 2000;Van Steenbergen & Band, 2013). In addition, neuroimaging, including computed tomography (CT) and magnetic resonance imaging (MRI), has been used to correlate changes in pupillary functions with cerebral lesions in patients with stroke and other brain disorders (Peinkhofer et al., 2018). In the same vein, electrical stimulation of cortical areas such as the frontal eye field (Brodmann area 8) has been investigated to correlate pupillary and cortical function in non-human primates (Becket Ebitz & Moore, 2017).
Although pupillary function is of considerable interest to neurologists, ophthalmologists, neuroscientists, physiologists and psychologists, the exact mechanisms of supratentorial modulation of pupillary function remains poorly understood. Previous (unsystematic) reviews have focused mainly on cognitive aspects such as attention but not on pupillary cortical control (Laeng, Sirois & Gredebäck, 2012;Granholm & Steinhauer, 2004;Van der Wel & Van Steenbergen, 2018).
Therefore, in this review we aimed to identify (a) the cortical and subcortical areas and (b) the behavior and cognitive processes that modulate pupillary function in humans and non-human primates.
questions we included non-human primates with or without cerebral lesions studied to correlate pupillary function with cerebral cortical and/or subcortical gray matter areas and with specific cognitive or emotional cerebral processing related to experimental invasive or non-invasive stimulation. For exclusion criteria, the reader is referred to the protocol review (File S1).

Outcome measures
The main outcome measure was a change in pupillary function, i.e., either a variation of the pupillary diameter or a difference in the light reflex (e.g., a longer latency period), compared to a baseline value or a control group.

Index tests and interventions
The index tests comprised neuroimaging (CT, MRI including functional MRI, PET, SPECT), post-mortem examination revealing the extent of brain lesions, quantitative pupillometry (Eye Link 1000 and similar devices) and visual inspection of pupillary function. Concerning interventions, we included all studies with invasive procedures such as electrical cortical and/or subcortical stimulation or induced cerebral lesions as well as non-invasive interventions such as cognitive and emotional tasks or sensorial stimulation of healthy humans, humans with specific cerebral lesions (see above) and non-human primates.

Electronic literature search strategy
We searched MEDLINE (PubMed), EMBASE and Scopus for relevant literature from January 1st, 1960 to November 15th, 2018. As a search strategy, we used both free textwords (TW) and controlled terms obtained with medical subject headings (MeSH). For search strategy and search terms refer to review protocol (File S1). Reference lists were manually screened for further relevant articles.

Selection of studies, data extraction and management
Titles and abstracts were first reviewed. Eligible studies were assessed on the basis of their full text and referenced using Mendeley Software (https://www.mendeley.com). Data were extracted by the first author and checked by the senior author. Preferred Reported Items for Systematic reviews and Meta-analyses (PRISMA) guidelines were followed (Liberati et al., 2009) (see File S2).

RESULTS
We screened 8809 papers in the primary search; three additional publications were manually added. After the exclusion of duplicates, studies with different topic and subjects below 18 years of age, 856 citations were screened for eligibility criteria on an abstract basis. Three-hundred and fifty-five articles were analyzed with a full text review, and 258 studies were included for the final analysis. Figure 1 provides

PICO 1: Pupillary changes associated with cortical lesions in humans
Cerebral areas that may modulate the pupillary light reflex were examined in three studies involving patients with cerebrovascular lesions. One study assessed pupillary dilatation as an index of arousal and reward processing during an oculomotor capture task (Manohar et al., 2016), revealing diminished pupillary dilatation in patients with chronic ventromedial prefrontal damage (Brodmann areas 11 and 25) due to subarachnoid hemorrhage as compared to healthy controls. Another, retrospective study of patients with cerebrovascular lesions (Herman, 1975), showed persistent anisocoria associated with lesions involving the right or left middle cerebral artery (MCA) territory in the absence of oculomotor nerve compression, but neuroimaging was not available and study results should be cautiously interpreted. In contrast, ischemic stroke lesions were verified using CT in a recent, prospective study, in which investigators assessed how anterior circulation strokes involving the prefrontal eye field (Brodmann area 8) and/or the insular cortex (Brodmann areas 13 and 16) affected pupillary function. Patients with strategic infarcts in either of these areas showed subtle differences during the dilatation phase of the pupillary light reflex, but not patients with infarcts in other cerebral regions or neurologically normal controls (Peinkhofer et al., 2018).

PICO 2: Pupillary changes associated with cognitive and emotional activity in humans
Most of the papers (n = 242) referred to changes in pupillary diameter during cognitive and/or emotional processes in humans. One hundred eighty-one (75%) assessed pupillary diameter as an index of mental effort during different cognitive activities. Sixty-one studies (25%) focused on the relationship between emotional arousal and pupillary reaction (Table 1).

Pupillary dilatation: cognition
Several studies recorded pupillary dilatation with memory tests, revealing how a change in diameter is related to memory retrieval. Pupillary dilatation occurred during testing Table 1 Human studies of the influence of cognitive and emotional processes on pupillary function. Every study is categorized depending on the specific task required and/or type of stimuli used (first column on the left ) and on the observed pupillary response (central and right column).

Pupillary Dilation
Pupillary responses other than dilation

PICO 3: Pupillary changes associated with cortical stimulation and lesions in non-human primates
Pupillary dilatation occurs in non-human primates in response to electrical stimulation of the frontal eye field (Brodmann area 8) during passive viewing tasks (Lehmann & Corneil, 2016) (''probe in, probe out'' conditions (Becket Ebitz & Moore, 2017)), and of the superior colliculus (Wang et al., 2012;Joshi et al., 2016) during passive fixation tasks. One study compared non-human primates with amygdala lesions to healthy controls during a free viewing task; pupillary dilation was similar in both groups, but the pupillary light reflex was diminished in the lesion group (Dal Monte et al., 2015) (Table 2).

DISCUSSION
This systematic review reveals that pupils do not only dilate and constrict in response to light, but a large number of cognitive and emotional processes affects pupillary function and leads to pupillary dilatation and, less often, to constriction (Table 1). Pupil diameter may serve as an index of brain activity, reflecting mental efforts (or lack of efforts). Thus, our pupils dilate, when we are focused in contrast to when we let our minds wander; they dilate when we are dishonest and lying; when we enjoy or dislike what we are seeing; and when we are engaged in learning and processing of information.

Pupillary dilatation
The most commonly observed response following emotional or cognitive tasks is pupillary dilatation. In humans, as well as in non-human primates, this is due to sympathetic Table 3 Non-human primate studies on the relationship of cognitive and emotional processes with pupillary function and activation of cortical/subcortical areas.
Characteristics of studies investigating which tasks and/or sensorial stimulus evoked a pupillary response and which cerebral areas were simultaneously activated. activation or parasympathetic inhibition or a combination of the two (Steinhauer, Condray & Kasparek, 2000) and based on unconscious mechanisms. Hence, tasks that require a high amount of attention such as memory retrieval, mental arithmetic or language processing elicit a sympathetic activation. Similarly, emotional sounds and images induce a state of arousal, which involves sympathetic activity leading to pupillary dilatation. Cerebral structures involved in vigilance, arousal and attention and responsible for changes in pupillary diameter during cognitive and emotional processes include the locus coeruleus (Joshi et al., 2016;Murphy et al., 2014), the superior colliculus (Wang et al., 2012) and multiple regions of the frontal/prefrontal cortex (Brodmann areas 8, 9 and 11 ) (Becket Ebitz & Moore, 2017;Lehmann & Corneil, 2016) (Fig. 2). Of these, the locus coeruleus seems to be the most influential mediator of the pupillary light reflex. This pontine nucleus is part of the ascending reticular activating system (ARAS) and intimately and reciprocally linked to the orbitofrontal cortex (Brodmann area 11) and the anterior cingulate cortex (Brodmann area 24 and 32) (Johansson et al., 2018;Geva et al., 2013) which are both fundamental to motivational relevance and target fixation. Evidence from studies of these networks supports the notion that attention and vigilance are related to the regulation of pupillary light reflex. Thus, the locus coeruleus modulates an excitatory connection to the sympathetic network of the pupil (in particular to the intermediate-medial-lateral cell column of the spinal cord) and an inhibitory connection to the parasympathetic pathway (directing to the Edinger Westphal nucleus). Activation of the locus coeruleus leads to increased sympathetic and decreased parasympathetic activity and, consequently, pupil dilatation (Samuels & Szabadi, 2008). Two recent studies highlight these aspects. According to Joshi et al. (2016), the locus coeruleus acts together with the inferior and superior colliculi, as well as the anterior and posterior cingulate cortex (Brodmann areas 23, 24, 31, 32) likely in response to increased vigilance and alertness, thereby modifying the pupillary diameter. The second study (Schneider et al., 2016), conducted on human beings, confirms this theory and shows that, based on data from resting state magnetic resonance imaging, pupil dilatation is related to an increased activity of the thalamus and frontoparietal regions (Brodmann areas 6, 39, 40), involved in the so-called tonic alert status and vigilance, and to increased metabolism of the visual and sensory-motor regions.
Besides the locus coeruleus, the superior colliculus seems to play a key role in modulating the pupillary light reflex. Wang et al. (2012) reported that pupils temporarily dilate after stimulation of the intermediate layer of the superior colliculus in non-human primates. Further, Mill, O'Connor & Dobbins (2016) and Herweg, Sommer & Bunzeck (2017) suggested that the superior colliculus receives neuronal inputs from temporal, frontal and parietal areas and basal ganglia, especially striatal and pallidal neuronal groups, leading to pupillary dilatation associated with memory tasks.
In addition, different experimental conditions in macaques show that stimulation of the frontal eye field (Brodmann area 8) might modulate the pupillary light reflex (Becket Ebitz & Moore, 2017;Hampson, Opris & Deadwyler, 2010). For instance, simultaneous micro-stimulations of the frontal eye field and of pretectum structures enhance the activity of frontal eye field neurons with similar spatial tuning and reduce, or Figure 2 Schematic representation of pupillary pathways that are activated during cognitive and emotional processes, including arousal and vigilance. Pathways, connecting the cortical areas to the parasympathetic system and the sympathetic system, are inhibitory or activating. Neurons emerging from the locus coeruleus inhibit the parasympathetic system at the Edinger Westphal nucleus and activate the sympathetic system via connection to the spinal cord tract of the sympathetic system. Red arrows: connections from cortical areas involved in the autonomic control i.e., anterior/posterior cingulate cortex and insular cortex. Blue arrows: connections from other cortical areas involved in visual processes. Green arrows: connections from subcortical structures i.e., locus coeruleus and superior colliculus. For reference to Brodmann areas, see text.
Full-size DOI: 10.7717/peerj.6882/ fig-2 even suppress, the activity of neurons with different tuning (Schlag, Dassonville & Schlag-Rey, 1998). From these observations, Becket Ebitz & Moore (2017) hypothesized that the frontal eye field and parts of the pretectum interact in regulating pupillary function. Although evaluation of the pupils is part of the routine clinical examination, only few human studies have correlated pupillary function with specific cerebral areas to replicate results from (invasive) non-human primate studies. Systematic studies on pupil diameter have been conducted in three clinical settings: Raised intracranial pressure, which may lead to oculomotor nerve compression and brain herniation; traumatic brain injury and cerebrovascular disease; but only studies on the latter have provided data on candidate cerebral areas that may regulate the pupillary light reflex. The classical work on this topic is by Herman (1975). In a retrospective study of 363 cerebral infarction patients, having excluded previous ocular pathology, local trauma, and active blood serology, the author reported that 5% of the examined patients had an asymmetrical pupillary response. Among the patients with pupillary asymmetry, 80% showed contralateral hemispheric stroke lesions, associated with other focal neurological signs and 20% of the patients had a dilated pupil homolaterally to the hemispheric lesion. A more recent work (Peinkhofer et al., 2018) found differences in the second phase of the pupillary light reflex, i.e., when pupils dilate back to baseline diameter, in patients with prefrontal eye field and/or insular infarcts (Brodmann areas 8,13 and 16). In this study the authors assessed pupillary function in patients with an acute anterior circulation stroke, treated with endovascular thrombectomy, and compared patients with infarcts in the prefrontal eye field and/or insular cortex to patients with infarcts in other areas (based on neuroimaging). No difference was found in the overall pupillary function, but subtle changes were observed in the dilatation phase. Therefore, the prefrontal eye field and/or insular cortex may have a role in modulation of pupillary light reflex, influencing the autonomic system directly or indirectly, perhaps via connections to subcortical structures such as the locus coeruleus. Similarly, it seems that subjects with focal damage in ventral and medial prefrontal cortex (Brodmann areas 11 and 25) have a constant reduction of reward-induced autonomic pupil responses, compared to age-matched, healthy controls, confirming the involvement of these areas in the cortical modulation of pupillary light reflex (Manohar et al., 2016).

Pupillary constriction
Pupillary constriction, induced by the parasympathetic system, is less frequently associated with cognitive or emotional processes than pupillary dilatation. It can be related to parasympathetic eliciting emotions such as disgust (Ayzenberg, Hickey & Lourenco, 2018) or memory tasks (Naber et al., 2013). The latter result is in contrast to the great majority of the studies on this topic (Kafkas & Montaldi, 2015;Kafkas & Montaldi, 2012;Bradley & Lang, 2015) that reveal pupillary dilatation but the difference seems to be mainly methodological, that is, related to the temporal evolution of the pupillary reflects analyzed: the first phase (i.e., constriction) or the second phase (i.e., dilatation), which are present in any task involving visual information processing.

Limitations
It should be noted that this systematic review has some limitations. First, we excluded studies with less than 15 patients, perhaps missing some relevant research. Second, the tools used to measure pupillary function were not the same across studies and, third, the exclusion criteria regarding previous neurological or ocular pathologies were not always clearly stated. Finally, it should be noted that pupillary function can be influenced by medication affecting the noradrenergic system, and very few papers provided information about the presence of absence of such medication. On the positive side, this paper is the only recent review on the topic and includes more than 200 publications on cortical pathways and behaviors modulating pupillary function.
In summary, this review shows that: -cognitive efforts and processing of emotional stimuli influence pupillary diameter in both humans and rhesus macaques, typically evoking pupillary dilatation; -pupillary constriction occurs in response to light stimuli, both real and imagined, suggesting a cortical influence on subcortical reflex pathway; -damage to salient cortical and subcortical areas such as frontal and prefrontal cortex, as well as key structures for autonomic control, seem to affect pupillary function by modulating the pupillary diameter; -and micro stimulation of the frontal eye field (Brodmann area 8), locus coeruleus and superior colliculus in non-human primates leads to pupillary dilatation, suggesting involvement of these areas in the pupillary light reflex.

CONCLUSIONS
Cognitive and emotional processes evoke a change in pupillary diameter, typically dilatation, in both humans and non-human primates, reflecting vigilance, arousal or attention. Stimuli related to light, whether real or imagined, elicit a pupillary constriction. Both dilatation and constriction are dependent on autonomic activation with cortical influence. The main structures involved are the locus coeruleus and the superior colliculus because of their direct and indirect connections to the Edinger-Westphal nucleus. Furthermore, cortical areas such as the prefrontal and the frontal cortex, particularly the frontal eye field (Brodmann area 8) and areas involved in autonomic control, such as insular cortex (Brodmann areas 13 and 16) and anterior cingulate cortex (Brodmann areas 24 and 32), modulate the pupillary light reflex via connections to subcortical structures and the Edinger-Westphal nucleus.

ADDITIONAL INFORMATION AND DECLARATIONS Funding
The authors received no funding for this work.

Competing Interests
The authors declare there are no competing interests.

Author Contributions
• Costanza Peinkhofer performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft.
• Gitte M. Knudsen authored or reviewed drafts of the paper, approved the final draft.
• Rita Moretti performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, authored or reviewed drafts of the paper, approved the final draft.
• Daniel Kondziella conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft.

Data Availability
The following information was supplied regarding data availability: No raw data were generated; this is a systematic review.

Supplemental Information
Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.6882#supplemental-information.