Mental health disorders (MHDs) describe enduring altered cognition, emotion, and behavior and remain a leading cause of the global burden of diseases. Recently, the locus coeruleus norepinephrine system (LC-NE) has been established as a modulator of sensory processing that translates acute stress responses into neuronal excitability. We propose this LC-NE functioning to explain a differential susceptibility to psychosocial adversity that underlies dispositional risks for MHD (Figure 1). LOCUS-MENTAL will put this to the test. We will assess LC-NE functioning in preschoolers as a predictor of later psychopathology. We will use pupillometry as a feasible neuroimaging technique to quantify LC-NE functioning in preschool children. LOCUS-MENTAL ultimately aims to generate an objective tool of individual risk prediction that informs a targeted prevention of lifetime MHD.

LOCUS-MENTAL uses a transdiagnostic, dimensional, and developmental approach to investigate LC-NE functioning as a predictor of MHD. Variations of LC-NE functioning have been associated with phenotypes of psychopathology [1-3]. Our work validated pupillometry as a feasible method to assess LC-NE functioning, which will now be scaled to the application in an outpatient clinic for early detection in the general population [4,5]. Preschool children will be assessed in an accelerated longitudinal study to reveal prospective effects of LC-NE functioning on psychopathology in the school age. These key concepts are discussed below.

Comorbidity describes the co-occurrence of more than one diagnosis of MHD. Comorbidity is common, associated with more impairment and worse lifetime outcomes [6]. We lack the evidence to effectively treat MHD comorbidity at the time of this writing [7]. This may originate from conceiving MHD as independent diagnoses in clinical classification systems, which emphasized a focus on isolated diagnoses. In contrast, transdiagnostic psychopathology uses MHD comorbidity patterns to derive overarching factors. A bi-factor model with 3 factors showed the best fit in children [8]. An internalizing factor describes problems within individuals (depression and anxiety), whereas an externalizing factor describes problems between individuals (conduct problems and attention deficit hyperactivity disorder). Positive correlations between internalizing and externalizing established a superordinate p-factor that describes a liability for MHD. These 3 factors are continuous dimensions, in which manifestations below clinical thresholds are of prognostic relevance [9]. This transdiagnostic and dimensional perspective is expected to expose etiological mechanisms across MHD [10].

Comorbidity of MHD may further represent an endpoint where a secondary diagnosis arises because of a preceding primary diagnosis. This sequential comorbidity emphasizes a developmental perspective on MHD [11]. Most adults presenting comorbidity received an initial diagnosis in childhood, and half of all initial diagnoses occur before adolescence [6]. The preschool age may provide a sensible phase to identify mechanisms that influence developmental trajectories of lifetime psychopathology [12]. In addition, preschoolers differ in their risk for psychopathology even for shared psychosocial adversity [13]. This has been attributed to a differential susceptibility that explains an interaction of dispositional risks and stress [14]. We aim to identify a neurobiological mechanism of differential susceptibility [15] that mediates the impact of stress and contributes to the development of psychopathology.

The LC-NE system is a neuromodulatory network orchestrating the brain’s response to sensory stimuli [16]. The LC-NE comprises the bilateral pontine locus coeruleus (LC) and its projections, which is the primary source of cerebral norepinephrine. Neuroimaging has suggested a bidirectional connectivity of the LC with the anterior cingulate cortex (ACC), thalamus, cerebellum, and the temporoparietal junction [17]. Compared to the substantia nigra, another neuromodulatory nucleus, the LC has shown increased connectivity to the visual, parietal, and motor cortices and the cerebellum [18]. This suggests a relative focus of the LC in earlier stages of sensory processing that perpetuates cognition and behavior.

The Adaptive Gain Theory (AGT) describes LC-NE functioning in cognition and behavior as a performance optimizer with 2 interchanging modes [19]. An exploration mode describes LC-NE tonic activity with a variable low frequency (1‐5 Hz) that corresponds to arousal levels [20]. In contrast, an exploitation mode describes LC-NE phasic activity with transient frequency and amplitude bursts that enhance the sensory processing of task-relevant stimuli [21]. At the system level, LC-NE activity reflects alerting attention [22]. At the neuron level, this has been elaborated to a mechanism of sensory selectivity. LC-NE tonic and phasic activity release norepinephrine which causes an autoinhibition of spontaneous neuronal activity and an excitation of input-driven activity [23]. This specific mechanism of neuronal gain modulation emphasizes a sensory processing of salience [24]. In the glutamate-emphasizes-noradrenergic-effects model [25], a salience is neurochemically indicated by a release of glutamate and can be independent of top-down control [24]. Under relative arousal, LC-NE activity initiates a positive feedback loop of glutamate and norepinephrine interactions that leads to excitatory hot spots of functional activity and biases large-scale networks [26] toward sensory processing of salient stimuli [27]. It has been observed that, in mice, a moderate stimulation of LC-NE activity enhanced thalamic feature selectivity, increased information transmission, and led to improved perceptual performance [28]. LC-NE functioning has been established as a key modulator of arousal and sensory processing with subsequent effects on perception and behavior [16,29].

LC-NE activity causes a dilation of the eye’s pupil. This pathway is indirect, as the LC-NE inhibits the Edinger-Westphal nucleus that typically emphasizes pupil constriction [30]. Pupil dilation has historically been used as an indicator of arousal and cognitive load [31]. Recently, pupil dilation has been differentiated into a baseline pupil size (BPS) and a stimulus-evoked pupillary response (SEPR), which index changes in LC-NE tonic and phasic activity, respectively [32]. These indices are assessed by pupillometry in video-based eye tracking. A study of combined functional magnetic resonance imaging and pupillometry showed a correlation of the LC’s signal with pupil dilation during the resting state (BPS) and with pupil dilation to oddball stimuli (SEPR) [33]. In a rodent model, an LC-NE stimulation predicted pupil dilation after 200‐300 milliseconds [34]. Other subcortical regions elicited pupil dilation with larger delays, which indicated an indirect activation of the LC [35]. In an optogenetic mouse model, pupil dilation marked relative changes in LC-NE activity, whereas variability challenges pupil size as a readout of absolute LC-NE activity [36]. This can be addressed by the application of baseline-corrected metrics [37]. In humans, the assessment of LC-NE activity is difficult with magnetic resonance imaging and electroencephalography (EEG) due to the LC’s location and small size [38]. This is pronounced in children that often induce more movement artifacts. Pupillometry now provides an index of LC-NE tonic and phasic activity (BPS and SEPR) that can be reliably measured with eye tracking. This innovative neuroimaging allows us to assess LC-NE functioning even in preschoolers in a clinical setting [5].

Psychopathology often develops by an interaction of dispositional risks and stress [39]. Dispositional risks describe internal features that increase a latent burden. In children, stress often refers to adverse childhood experiences like neglect and further extends to psychosocial domains like peer victimization [40]. If sufficient stress occurs upon dispositional risks, a psychopathology may develop. The preschool age is sensitive to this interaction and describes a precursor phase of lifetime psychopathology [15]. In preschoolers, dispositional risk profiles are observed, which map to polygenic scores for MHD that develop much later [11]. We propose that the LC-NE functioning in preschoolers underlies dispositional risks and shapes risk trajectories across diagnoses.

Stress like adverse childhood experiences is the other important determinant of psychopathology. An acute stress response is an internal coping mechanism to stress triggered by a neuronal network including the amygdala [57]. First, an autonomic response increases sympathetic activation for a momentary fight-flight-or-freeze response. A hypothalamic-pituitary-adrenal (HPA) axis then releases glucocorticoids like cortisol for sustained physical readiness. This endocrine stress response initially supports stress coping but cumulatively upregulates the HPA axis, which ultimately wears the body [58]. It explains the negative impact of stress over time. However, the acute stress response is also translated to a neuronal excitability that shapes our immediate perception in the face of stress [59].

This immediate neuro-translation of stress is moderated by LC-NE functioning [60]. The LC-NE has been established to initiate the autonomic stress response [61]. In addition, the HPA axis expresses a corticotropin-releasing hormone that binds to LC receptors and upregulates LC-NE tonic activity early in the endocrine stress response [62]. The upregulation of LC-NE tonic activity attenuates LC-NE phasic activity, which increases stimulus reactivity at the expense of sensory selectivity [63]. It reflects hyper-arousal [64] that intensifies the impact of stress [13]. For example, a liability to social stress in rodents was explained by an LC-NE modulation [65]. In humans, anxiety and stress-related disorders have been associated with increased LC-NE tonic activity [53,66]. Excessive LC-NE tonic activity in response to stress-inducing adversity might be an etiological mechanism of anxiety symptoms [62,66] that warrants an empirical evaluation in longitudinal studies.

In addition, LC-NE phasic activity is involved in the stress response by an LC projection to the amygdala in a subliminal response to fear [67]. LC-NE phasic activity emphasizes a neurophysiological response to emotional stress [68]. This has been explained by a different distribution of adrenergic receptors between the prefrontal cortex and the amygdala, where LC-NE phasic activity causes an inhibition of executive control and an excitation of the limbic fear response [69]. It has been observed in mice, pharmacological blocking of LC-NE phasic activity reduced amygdala activity during fear conditioning [70]. In a seminal study with healthy humans, LC-NE phasic responsivity and functional coupling with the amygdala during a neurocognitive task predicted internalizing symptoms following stress-inducing life events [2]. Accordingly, posttraumatic stress disorder has been associated with elevated LC-NE phasic activity in response to loud noises [71]. This supports increased LC-NE phasic activity as an amplifier of the neurophysiological fear response in the face of stress.

LC-NE activity in stress regulation can, thus, be described as a (neuro-)translation of the acute stress response to neuronal excitability [72]. We propose that LC-NE functioning as a mechanism in young children that shapes a subjective perception of stress. Distinct LC-NE functioning profiles may emphasize psychological states (eg, threat perception) that increase the psychopathological impact of stress. Thus, interindividual differences in LC-NE functioning are expected to mediate the negative impact of childhood adversity. We expect that elevated LC-NE tonic activity corresponds to hyperarousal under stress that represents a liability to transdiagnostic psychopathology (p-factor). We further expect deviant LC-NE phasic activity to differentiate the development of internalizing versus externalizing psychopathology in response to childhood adversity. Elevated LC-NE tonic and phasic activity have been associated with internalizing symptoms in adults (eg, anxiety and negative affect), whereas blunted LC-NE phasic activity might indicate a subtype at risk for conduct problems [73]. LOCUS-MENTAL will investigate LC-NE functioning in preschoolers that mediates a differential susceptibility to adversity and differentiates future profiles of transdiagnostic psychopathology (see Figure 1).

The primary aim of LOCUS-MENTAL is to establish pupillometric measures of LC-NE functioning as a predictor of transdiagnostic psychopathology in childhood. We hypothesize that LC-NE functioning represents a fundamental neurobiological mechanism in the development of MHD that could serve as a biomarker for early risk detection. Transdiagnostic psychopathology is assessed with the caregiver-reported Child Behavior Checklist (CBCL), where clinical relevance is defined as reaching established cutoffs (T>60). LC-NE functioning is assessed with pupillometry. The study outcome could improve targeted prevention and intervention for MHD comorbidity in children. To achieve this aim, LOCUS-MENTAL is structured into 4 consecutive objectives, each with specific tasks within the following project plan (Figure 3).

LOCUS-MENTAL will achieve its objectives by combining pupillometric indices of LC-NE functioning and using a prospective panel study design (Figure 4). Pupillometry is assessed by video-based eye tracking during basic paradigms that are feasible in preschoolers. Transdiagnostic psychopathology is assessed by CBCL externalizing, internalizing, and total T-scores. Childhood adversity is assessed by interviewing caregivers, children, and clinicians. Pupillometric indices, transdiagnostic psychopathology, and childhood adversity are repeatedly measured in an accelerated longitudinal study: We assess preschoolers (aged 4‐6 years) in an initial assessment with reassessments after one year (aged 5‐7 years) and after 2 years (aged 6‐8 years). This accelerated design allows revealing causative effects of LC-NE functioning on psychopathology from 4 to 8 years of age. In addition, this allows quantifying LC-NE functioning as a mediator in the effect of childhood adversity on psychopathology. In distinct subtasks, LC-NE functioning is elaborated as a neurophysiological mechanism of established dispositional risks, while LC-NE functioning in stress regulation is explored by combined pupillometry and analysis of hair cortisol levels and salivary cortisol after a social stress test. This will characterize LC-NE functioning profiles in the interaction of dispositional risk and stress. Finally, LOCUS-MENTAL applies normative modeling on the pupillometric data to develop a tool for risk prediction of developmental trajectories in transdiagnostic psychopathology. An overview of all measures is provided below (Table 1). The reporting of this study protocol follows the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for longitudinal cohort studies [81]. The completed checklist is available in Checklist 1.

The main recruitment avenues will be the outpatient clinic for early detection at the Department for Child and Adolescent Psychiatry at the Goethe University Hospital. In the help-seeking population, externalizing disorders might be overrepresented, which is why we will overrecruit preschoolers with internalizing symptoms. We aim for a representative sample by recruitment across all levels of psychopathology. This is why we will also recruit in the general population by distributing flyers with study information at local health care and public institutions, pediatricians, and day care facilities and promoting the study in targeted social media groups and channels. Exclusion criteria will be known genetic syndromes and neurological diseases. Caregivers will receive a summary report of their child’s assessment (eg, IQ evaluation) upon completion of each assessment wave.

The experimental assessments will be carried out at our combined eye-tracking and EEG lab that we established at the clinic. The procedure is framed as a play experience. The child watches a sequence of their favorite movie on the stimulus presentation screen. This is followed by the assessments where the child may sit on their caregiver’s lap. Each paradigm is designed to enable breaks. Questionnaire measures may be filled out by the caregivers via our in-house online tool to decrease the participant burden, whereas structured clinical interviews and child questionnaires will be carried out by trained clinical staff.

Sample sizes (n) were determined based on recruitment feasibility and the ability to detect moderate effects with a power of β>.80. For each task, we estimated the power by Monte Carlo simulations (k=1000) within respective models [97,98]. In the following sections, the sample size and power are reported, respectively, for each specific objective or task.

BPS and SEPR are assessed for each trial in each task. BPS is defined as the pupil size prior to stimulus onset, while SEPR represents a pupil size change in response to stimuli. BPS and SEPR will be evaluated in the auditory oddball task in response to standard trials as neurophysiological habituation [99] and to oddball trials as sensory selectivity [33].

To assess the verbal and cognitive abilities, we will conduct 2 subtests from the Wechsler Preschool and Primary Scale of Intelligence (WPPSI) [93].

We will validate the pupillometric indices by a multitrait-multimethod design. In a homo-method analysis, the BPS and SEPR during the auditory oddball task (Figure 5A) will be compared to BPS and SEPR in tasks of LC-NE functioning with irregularity in rapid sound sequences [24] (Figure 5B), a visual oddball task [91] (Figure 5C), and an audio-cued visual search [92] (Figure 5D).

In a hetero-method analysis, the BPS and SEPR during the auditory oddball task will be compared to event-related potentials (ERPs) in concurrent EEG. We investigate oddball ERPs that are likely associated with LC-NE activity (P300, MMN) [27,100]. Finally, the pupillometric measures will be compared to divergent measures of cognitive and language ability as assessed by WPPSI-IV [93]. Patterns of high intercorrelations will be used to create a concise pupillometric battery of LC-NE functioning (~15 min). BPS and SEPR are reliably assessed in few oddball trials (k=20).

A successful task manipulation will be indicated by a higher SEPR in oddball versus standard trials (ie, sensory selectivity) and declining BPS across standard trials (ie, habituation). We expect convergent validity with moderate-to-high correlations between the auditory oddball SEPR and the pupillometric measures in the validation tasks (including SEPR to a change from regularity to irregularity, SEPR to visual oddball stimuli, and SEPR before target detection). We expect convergent validity in the hetero-method analyses, whereas we expect divergent validity with cognitive and language development.

We only include paradigms that are feasible in young children. This shall ensure a sufficient completion rate in preschoolers (>80%). The concurrent presentation of visual stimuli to maintain attention during the auditory oddball task may impede the SEPR. However, we recently showed in independent data that visual presentation during an auditory oddball task still induces distinct task-evoked pupillary responses.

The study was approved by the Ethical Committee (reference number: 2024‐2160) and the Data Protection Officer of the Medical Faculty of Goethe University Frankfurt. To date, no adverse events have been reported by any comparable study, and thus we expect no risks nor harm to our participants. The study is planned and will be conducted and analyzed according to Good Scientific Practice (GSP) guidelines, the Declaration of Helsinki, and the European Data Protection Directive (DSGVO). Study participation is on a voluntary basis, and participants can withdraw from study participation at any time without any adverse consequences concerning their medical or psychiatric treatment in our clinic. Participants will also receive monetary compensation for their time. All relevant information regarding data protection and the voluntary basis of participation will be communicated to participants and their legal caregivers verbally, as well as in the form of written informed consent forms. We also obtain verbal assent from the preschool participants. The child’s right to withdraw at any time will be prioritized over parental consent. During assessments, the child will be closely monitored by the experimenter that is present at all times, and signs of acute distress (such as crying, physical withdrawal, or verbal expression of a desire to stop) will lead to termination of the assessment. In cases of suspected or disclosed adversity, the research team will inform the principal investigator immediately (licensed psychotherapist for children), who offers a voluntary counseling appointment to discuss potential adversity with the family and the child. In line with German child protection laws and in accordance with our clinical guidelines, the quantity and quality of suspected adversity might result in further support by our department or even a report to authorities (eg, imminent danger of ongoing adversity). We acknowledge that the development of a risk assessment tool could potentially lead to stigmatization and misuse. With our data management plan outlined below, we made sure that individual risk prediction cannot be associated with personally identifiable information. We further refrain from any nonscientific use of the prognostic risk marker. The rights to the information of prognostic risk remain with the participants in accordance with DSGVO. We are confident that a prognostic risk marker also provides the opportunity to shift from a health care system that focuses on intervention after manifestation of psychopathology to a focus on targeted prevention. Children at an elevated risk for mental health problems as identified by LOCUS-MENTAL could receive further case management at our outpatient clinic, where we decide on useful next steps on a case-by-case basis with evidence-based treatments. For example, caregivers of children at risk could be offered the opportunity to receive parent trainings that are established in the treatment of mental health problems in children. This may aid in the early prevention of the initial manifestation of psychopathology. Our newly established outpatient clinic for early detection could be the ideal environment to use risk assessment information for positive clinical outcomes.

LOCUS-MENTAL applies a noninvasive neuroimaging methodology that allows indexing neuronal activity by video-based pupillometry. Previous work showed feasibility and established this methodology in clinical samples of children. In addition, we apply further noninvasive assessments with electroencephalography. These measures introduce a burden for the participants and can induce a state of irritability but do no harm. This also applies to the application of an age-appropriate social stress test. The test induces a temporary impression that the child’s performance was insufficient in a task, which has been associated with a cortisol release response that needs to be quantified to achieve the objective of task 3.2. The social stress test thus includes a comprehensive disclosure strategy after the test to alleviate any negative effects. In task 3.2, we further collect tissue as hair samples (cutting off hair strands) and saliva samples (chewing an artificial fiber). These methods of tissue collection are also entirely non-invasive. The tissue samples will be processed by experienced contractors and subsequently destroyed. In the project, we further assess personally identifiable information that is handled according to the descriptions in section 2.4. We expect that this study does not involve an immediate risk of yielding knowledge, products, or technology that could intentionally be misused to cause substantial harm.

Questionnaire and interview data will be collected via our on-premises online tool according to the European General Data Protection Regulation (GDPR). We further do video-based eye tracking, apply EEG, and collect biological samples. Hair samples will be analyzed by Dresden Labservice with online solid-phase extraction in liquid chromatography-tandem mass spectrometry. Saliva samples will be analyzed by Daacro Contract Research with the ELISA Salimetrics Assay in a double estimation. Metadata will be generated by a printed case report form. Data are transferred to electronic databases that allow for automated quality controls and sanity checks. Eye tracking and EEG data will also receive quality control by peer-reviewed preprocessing pipelines. We established standard operating procedures to document research data, personnel will be trained in GSP, and all data handling will be in accordance with the GDPR. We differentiate between person-identifying information and research data that are separated in different networks, both behind the clinic’s firewall. The data can only be combined by a person identification code that is restricted in access to authorized personnel and stored on the clinical network. This is separated from a pseudonymized research network. All databases run on a Microsoft Access front- and back-end with an Active Directory system. Data are stored on on-premises servers with backups in a different building. After the project’s end, anonymized research data will be made available according to Open Science standards and archived for 10 years. The author (IT) will administer the project-specific databases. PhD students will have read and write rights with log protocols. A dedicated database manager with a permanent position is the administrator of all databases.

The LOCUS-MENTAL study uses targeted strategies across its work packages to manage missing data and ensure robust longitudinal findings. Work package 1 focuses on feasibility measures by using passive and child-friendly paradigms to achieve a projected completion rate of over 80%. Statistical analysis will exclude children that have per-task data validity rates of less than 50% or more than one missing task. The auditory oddball task is a marker task that is required to be included in the statistical analysis. We established standardized procedures of pupillometry data preprocessing to address within-trial missing data due to inattention. Work package 2 addresses high-burden EEG assessments by designating them as “more optional” in the context of the overall project, which allows for an expected 30% attrition rate without hindering subsequent study objectives. For the longitudinal assessments in work package 3, the study implements retention strategies like online data collection and locator documents. In addition, the statistical framework of work package 3 includes LMM and structural equation modeling to handle intermittent missingness across time points. Questionnaire data on an item level with per-measure missing rates up to 20% will be addressed by Multiple Imputation by Chained Equations (MICE) that preserves statistical power under the missing at random assumptions. However, children with increased irritability or increased burden may be more likely to drop out (ie, missing not at random or MNAR). Thus, pattern mixture models will be applied as a sensitivity analysis to determine if the developmental trajectories of LC-NE functioning are biased by specific attrition patterns. In addition, Inverse Probability Weighting (IPW) in the longitudinal data analysis can be used to adjust for non-random attrition by weighting the remaining participants based on baseline characteristics (eg, socioeconomic status or initial psychopathology scores) that predict dropout. Missing bio samples will be addressed by case-wise exclusion in the specific subanalysis. Finally, work package 4 uses v-fold validation to evaluate the predictive stability of the risk prediction tool.

The LOCUS-MENTAL protocol outlines a longitudinal study design with the primary aim to establish LC-NE functioning as a neurobiological predictor of later psychopathology in childhood. We anticipate that the validation phase (Objective 1) will demonstrate convergent and discriminant validity of the pupillometric markers BPS and SEPR with sufficient reliability. By developing and validating a child-friendly, pupillometry-based tool, this aims to bridge a critical gap between neurobiological mechanism and clinical risk prediction, which is based solely on behavioral observation. Accordingly, this is a prerequisite to ensure the method sensitivity and specificity. This would provide a concise test battery that assesses a neurobiological mechanism that could be implemented in future clinical studies capturing a potential biomarker of psychopathology in early childhood. Furthermore, we hypothesize that BPS and SEPR will cross-sectionally differentiate between transdiagnostic risk profiles, such as irritability and behavioral inhibition, and serve as a mediator between childhood adversity and psychopathology (Objective 2). This would provide an underlying neurobiological basis for temperament dimensions that are also based on caregiver reports. In the longitudinal data, as a core outcome, we expect to reveal baseline LC-NE functioning at age 4‐6 years to predict the development of psychopathology at school age (Objective 3). This central component of the LOCUS-MENTAL project would provide causative evidence for an impact of LC-NE functioning on transdiagnostic mental health conditions. These findings will be used in forming a normative model as a pupillometry-based risk prediction tool, establishing a key neurobiological susceptibility mechanism that mediates the impact of early adversity on psychopathology (Objective 4). This will allow a translation of the experimental findings to clinical practice. For example, children associated with an elevated risk for psychopathology based on LC-NE functioning could be supported by evidence-based treatments that are effective before the onset of a mental health condition. This would emphasize a targeted prevention approach. We conclude that the identification of a neurobiologically grounded risk marker could transform our mental health infrastructure from interventions that are applied years after the onset of mental health problems to early and targeted support. Ultimately, this aims to reduce the lifetime burden of mental health problems.

LOCUS-MENTAL focuses on a critical developmental window, the preschool period (aged 4‐6 years), a key time for the emergence of transdiagnostic risk factors, offering a unique opportunity for targeted prevention. This developmental period has been largely neglected in experimental research on lifetime mental health, as preschoolers have limited capabilities to follow task instructions. In this project, this is addressed by the application of feasible pupillometry as a primary method that is feasible in young and burdened children. The study provides a robust design, combining an initial multitrait-multimethod validation phase with a subsequent large-scale accelerated longitudinal study. This design strengthens internal validity through paradigm optimization and allows for examination of causal pathways and developmental trajectories over time.

The novelty of the paradigms for preschoolers is a limitation since established pupillometric paradigms for assessing specifically LC-NE functions have not been established in young children. The dedicated validation phase (working package 1) is specifically designed to address this issue. Another limiting factor is the longitudinal study design. This design comes with the risk of high participants’ attrition, which may affect the final statistical power. To address this, the study will use retention strategies (eg, regular contact, family-friendly appointment scheduling, and compensation) and will use appropriate statistical methods to handle missing data.

If the pupillometric markers are validated as successful predictors, future research could focus on clinical implementation. The natural next step would be the development of a pupillometry tool that can be applied outside the lab, for which head-mounted devices with built-in eye trackers are a promising avenue. Ultimately, this would deliver a child-friendly and low-cost automated screening tool for pediatricians or for educational settings.

The results of the study will be disseminated through peer-reviewed publications in high-impact journals. Findings will be presented at national and international conferences. Following Open Science principles, data analysis will be made available in online repositories, while deidentified data can be shared upon reasonable request in line with data privacy considerations. In addition, we engage in active patient and public involvement by media outreach and promoting our research on social media and our Mind-Child lab web page.