Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Sep 11;50(9):e70414. doi: 10.1111/acer.70414

Validity of an Online Cognitive Screening Tool for FASD: Testing the Equivalence of BRAIN‐online to Neuropsychological Assessment Measures

C R Veziris 1, M R Kelm 1, D Connely 1, G Chockalingham 2, J Wozniak 3, S N Mattson 1,✉
PMCID: PMC13569869  PMID: 42728660

ABSTRACT

Background

The diagnostic process for fetal alcohol spectrum disorders (FASD) requires investments in both time and cost. The Brief Assessment of Individual Neurobehavior, online version (BRAIN‐online), is a screening tool designed to identify cognitive and behavioral symptoms associated with FASD. BRAIN‐online is effective in distinguishing individuals with histories of prenatal alcohol exposure (PAE) and FASD; however, criterion validity of this screening tool has not been established. The goal of this study was to evaluate performance on BRAIN‐online against gold‐standard methods of measuring cognitive functioning (i.e., in‐person neuropsychological assessment).

Methods

Participants included 157 children (5–17 y) with (AE; n = 101) and without (CON; n = 56) PAE as part of the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD) Phases 4 and 5. Individuals completed BRAIN‐online and underwent a standardized in‐person neuropsychological assessment battery; caregivers completed the BRAIN‐online Behavior Questionnaire (BBQ) and a standardized behavioral questionnaire. BRAIN‐online measures were matched to in‐person measures, and scores were assessed for correlation. In addition, equivalence testing was performed on BRAIN‐online measures and their parallel in‐person measures. To assess clinical utility, neurobehavioral impairment as defined by clinical guidelines for diagnosing FASD was assessed using BRAIN‐online and in‐person neuropsychological assessment and compared to FASD diagnoses using sensitivity, specificity, and accuracy analyses.

Results

Seven of the eight BRAIN‐online scores were correlated with scores of their corresponding in‐person measures, and seven of eight BRAIN‐online subtests showed measurement equivalence with parallel in‐person neuropsychological assessment measures. BRAIN‐online and in‐person neuropsychological assessment had similar accuracy rates of classifying FASD, with BRAIN‐online having slightly higher sensitivity and slightly lower specificity.

Conclusions

BRAIN‐online subtests are correlated with and show measurement equivalence with gold‐standard measures of cognitive functioning. Therefore, BRAIN‐online may have clinical utility in identifying individuals who have symptoms associated with FASD and can benefit from comprehensive neuropsychological evaluation.

Keywords: equivalence testing (TOST), fetal alcohol spectrum disorders (FASD), neuropsychological assessment, online cognitive screening, prenatal alcohol exposure (PAE)

BRAIN‐online, an online screening tool for fetal alcohol spectrum disorders (FASD), was previously shown to accurately identify individuals with prenatal alcohol exposure. The current study demonstrated correlation and measurement equivalence to gold‐standard neuropsychological measures. BRAIN‐online and in‐person neuropsychological assessment also had similar accuracy rates of classifying FASD. In combination, these results indicate that BRAIN‐online may have clinical utility in identifying individuals who have symptoms associated with FASD and can benefit from comprehensive diagnostic evaluation.

graphic file with name ACER-50-0-g001.webp

Abbreviations

ADHD

Attention‐Deficit/Hyperactivity Disorder

ASD

Autism Spectrum Disorder

BASC‐3

Behavior Assessment Scale for Children, Third Edition

BBQ

BRAIN‐online Behavioral Questionnaire

BRAIN‐online

Brief Assessment of Individual Neurobehavior‐online version

CIFASD

Collaborative Initiative on Fetal Alcohol Spectrum Disorders

CVLT‐C

California Verbal Learning Test, Children's Version

D‐KEFS

Delis‐Kaplan Executive Functioning Scales

FASD

Fetal alcohol spectrum disorders

PAE

Prenatal alcohol exposure

TOST

Two One‐Sided Tests

WISC‐V

Wechsler Intelligence Scale for Children, Fifth Edition

Fetal alcohol spectrum disorders (FASD) are characterized by physical, cognitive, and behavioral symptoms associated with exposure to alcohol in utero and include fetal alcohol syndrome (FAS), partial fetal alcohol syndrome (PFAS), alcohol‐related neurodevelopmental disorder (ARND), alcohol‐related birth defects (ARBD), and neurobehavioral disorder associated with prenatal alcohol exposure (ND‐PAE; American Psychiatric Association 2022; Hoyme et al. 2016). Although not all FASD diagnoses require the presence of physical characteristics or confirmation of alcohol exposure in utero, with the exception of ARBD, all require the presence of measurably atypical cognition and behavior (Hoyme et al. 2016). FASD is thought to affect 1.1%–5% of children in the United States (May et al. 2018) and is more prevalent than other neurodevelopmental disorders such as autism spectrum disorder (ASD; Maenner et al. 2020). Thus, prenatal alcohol exposure (PAE) poses a significant public health concern with widespread implications for the well‐being of both individuals and their communities (Greenmyer et al. 2018).

Despite the high prevalence of PAE in the United States, rates of FASD misdiagnosis and underdiagnosis remain high, likely due to similarities in symptomatology to other neurodevelopmental disorders (e.g., attention‐deficit/hyperactivity disorder (ADHD) and stigma surrounding disclosure of PAE). Importantly, rates of PAE are estimated to be particularly high in populations such as children involved with the child welfare system and those with fewer socioeconomic resources (Sood et al. 2001). In fact, FASD is thought to affect an estimated 18.8% of children in the child welfare system (Astley et al. 2002; Engesether et al. 2024). Despite the prevalence of PAE in at‐risk populations, an estimated 80% of FASD diagnoses are routinely missed in foster care populations (Chasnoff et al. 2015), and even if individuals are identified as potentially having an FASD, wait times for a formal FASD diagnostic assessment can be up to 4.5 years (Popova et al. 2024). These gaps highlight the urgent need for FASD screening tools that are efficient, effective, and accessible given the high economic and public health burden of untreated FASD (Greenmyer et al. 2018).

Although there are commonalities between FASD and other neurodevelopmental disorders such as ADHD (e.g., difficulties with attention) and the two often co‐occur (Mattson et al. 2019), FASD presents a unique pattern of cognitive and behavioral difficulties that require specialized treatment. Cognitively, individuals may experience difficulties with executive functioning, learning and memory, language, visual–spatial abilities, and motor functioning (Mattson et al. 2019). Behaviorally, difficulties with attention, self‐regulation, and impulse control are common, leading to an increased rate of both diagnoses (Mattson et al. 2019). Together, these cognitive and behavioral challenges can interfere with daily functioning, academic success, relationships, and related outcomes across the lifespan (Mattson et al. 2019; Moore and Riley 2015). Incorrect or incomplete diagnosis may lead to less‐than‐optimal treatments; children with FASD often face challenges that cannot be treated with the same interventions as other neurodevelopmental disorders (Chasnoff et al. 2015). For example, research on the efficacy of pharmacological treatments tailored for ADHD in children with FASD is mixed (Ritfeld et al. 2022, 2024), suggesting that standard treatment for ADHD (psychostimulant medication) may not adequately address the broader needs of individuals with FASD. In addition, children with FASD are at an increased risk for environmental stressors such as involvement in foster care and adverse childhood experiences, which may exacerbate existing cognitive and behavioral challenges (Kautz‐Turnbull et al. 2023). Thus, accurate diagnosis leading to proper treatment is necessary to improve lifelong well‐being for individuals with FASD.

One significant barrier in the diagnostic process is that all diagnoses under the FASD umbrella include requirements for cognitive and/or behavioral impairment, necessitating an extensive evaluation (Coles et al. 2023; Hoyme et al. 2016; Wozniak et al. 2019). This evaluative process, commonly completed through in‐person neuropsychological assessment, requires both time and financial resources; neuropsychological evaluations can take anywhere from 1 to 8 h (Schaefer et al. 2023) and cost upward of $2000 on average in the United States (Kurniadi and Davis 2022). The expensive and time‐intensive nature of neuropsychological testing is a barrier to comprehensive assessment, particularly for children involved in the child welfare system and those with fewer economic resources, and can delay accurate diagnosis. In addition, inaccurate referrals may lead to an overwhelming number of assessments, especially with a limited number of FASD clinics, leading to long wait times for individuals who need FASD‐specific resources (Popova et al. 2024). This, in turn, increases risk for adverse outcomes, such as difficulties in school and trouble with the law (Rockhold et al. 2024). Early diagnosis of FASD is associated with a reduced risk of these adverse outcomes (Ritter et al. 2025), supporting the need for accurate, efficient, and accessible screening for FASD.

A tool that has furthered accessibility of cognitive screening is online neuropsychological testing, or tele‐assessment, which has allowed for efficient cognitive evaluation that increases the reach of clinical care through the removal of travel barriers. Online neuropsychological tests have been shown to be reliable and similarly as effective as in‐person neuropsychological testing in both pediatric and adult clinical populations across a variety of cognitive domains (e.g., language, IQ, memory; Alfano et al. 2024; Brown and Zakzanis 2025; Walker et al. 2023). Diagnostically, tele‐assessment has also been successfully used to assess for specific neurodevelopmental disorders such as ASD and ADHD in children (McNally Keehn et al. 2023; Ros‐Demarize et al. 2021). However, while online measures are effective in capturing cognitive functioning and remove some healthcare barriers (e.g., travel time and cost), they may still be time‐consuming and costly, as many still require the use of an examiner or psychometrist (Walker et al. 2023). In addition, these adapted neuropsychological evaluations may not comprehensively assess areas of difficulty that are necessary for an FASD diagnosis. Unlike other neurodevelopmental disorders that may be evaluated through caregiver and/or teacher report (e.g., ADHD; Ros‐Demarize et al. 2021), FASD necessitates a tailored neuropsychological battery and screening for specific cognitive and behavioral challenges. While tele‐assessment has proven to be a valuable tool, research remains limited on its use in FASD diagnostics.

The Brief Assessment of Individual Neurobehavior, online version (BRAIN‐online), is an online tool developed to screen for cognitive and behavioral challenges associated with FASD (Mattson et al. 2026). Designed to mirror traditional FASD neuropsychological evaluations, the purpose of BRAIN‐online is to identify cognitive and behavioral difficulties associated with FASD that indicate the need for further FASD assessment in a brief, online format at no or low cost and without the need for an examiner to be present. In doing so, BRAIN‐online has the potential to reduce the time and financial burden for families that are often associated with comprehensive diagnostic procedures. Although BRAIN‐online is not a diagnostic tool, the comprehensive yet specific approach of this assessment tool can greatly improve diagnostic efficiency for individuals with FASD by refining referrals for FASD diagnostic evaluations and, in turn, improving clinic wait times through its efficient screening method. In addition, this screening tool has the potential to reduce time in an in‐person neuropsychological assessment by efficiently screening for information about cognitive and behavioral difficulties, therefore having the potential to result in a truncated neuropsychological battery. Finally, the use of BRAIN‐online as a screening tool may eliminate the need for comprehensive FASD diagnostic assessment if performance is not consistent with FASD. The research on BRAIN‐online so far is promising; BRAIN‐online differentiates children and adolescents with PAE from children and adolescents without prenatal exposure with 80% accuracy in a research setting (Mattson et al. 2026). In addition, the behavioral component of BRAIN‐online showed excellent discrimination in differentiating those with PAE from those without, and its cognitive subtests were able to significantly predict group classification (prenatal alcohol exposure vs. no prenatal alcohol exposure), highlighting the ability of this measure to efficiently and accurately assess likelihood for an FASD.

Although BRAIN‐online has been shown to have high accuracy and construct validity (Mattson et al. 2026), the criterion validity of this tool has not yet been evaluated (i.e., validated against the gold‐standard neuropsychological assessments on which it was modeled). This is a crucial step in confirming that BRAIN‐online is indeed capturing the cognitive and behavioral domains of interest and thus will lead to appropriate diagnostic referrals and diagnostic accuracy. Therefore, the purpose of this study was to assess criterion validity for BRAIN‐online by evaluating (1) whether BRAIN‐online scores are comparable to scores received on similar in‐person neuropsychological assessments, (2) whether BRAIN‐online subtests measure domains of functioning as effectively as in‐person neuropsychological measures, and (3) whether BRAIN‐online can classify FASD‐risk as accurately as in‐person neuropsychological assessment. We hypothesized that BRAIN‐online would capture similar abnormalities, be equally as effective, and classify at‐risk individuals as accurately as in‐person neuropsychological measures, providing an accessible, efficient tool for FASD diagnostic screening.

1. Materials and Methods

1.1. Participants

Participants were 5–17‐year‐old youth (M = 9.3, SD = 3.23) and their caregivers from the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD) study, Phases 4 and 5 (2020–2025) from the Center for Behavioral Teratology at San Diego State University. Participants included children and adolescents (52.2% female) with moderate‐to‐heavy amounts of PAE (alcohol‐exposed group, AE; n = 101) and minimal to no amounts of PAE (control group, CON; n = 56). Per previous CIFASD criteria (Mattson et al. 2010), minimal exposure was defined as less than one drink per week on average and never more than two drinks on any occasion during pregnancy; moderate exposure was defined as one to 13 drinks per week on average or two to three drinks per occasion during pregnancy; heavy exposure was defined as exposure to an average of 14 or more drinks per week on average or four or more drinks per occasion during pregnancy. PAE information was collected via caregiver questionnaires and/or review of medical, legal, and social work records (see Mattson et al. (2010) and Jones et al. (2006) for more information about CIFASD methodology). As part of the larger CIFASD study, all participants completed BRAIN‐online at home and in‐person neuropsychological assessment. Participants in both groups were included if they spoke English and were not excluded if they had other mental health or neurodevelopmental concerns or disorders (e.g., ADHD, global intellectual impairment) or received special education services. Participants were excluded from participation for any of the following reasons: They experienced a significant head injury with loss of consciousness (> 30 min) or had a significant physical or psychiatric condition that interfered with their ability to complete study tasks. Data were excluded from analysis if they were missing information about PAE (n = 12), took more than 120 min to complete the cognitive subtests of the BRAIN‐online portion of the study (which is expected to only take 45 min on average; n = 13), or did not complete in‐person neuropsychological testing within a year of completing BRAIN‐online (n = 69). Our final sample size comprised 157 participants.

1.2. Measures

1.2.1. BRAIN‐Online

BRAIN‐online is an FASD screening tool developed to capture specific cognitive and behavioral domains associated with FASD. Tests were modeled after open‐source material or were developed to assess specific cognitive domains associated with FASD (Mattson et al. 2026). BRAIN‐online consists of seven subtests and one behavioral questionnaire that collectively measure: behaviors associated with FASD, fine motor speed, reaction time, inhibition, attention, memory, visual working memory, and executive functioning (see Table 1, for a description of the BRAIN‐online subtests). In addition, BRAIN‐online includes a demographics questionnaire, and both the demographic and behavioral questionnaires are filled out by the caregiver prior to BRAIN‐online completion by the child. All components of BRAIN‐online (i.e., demographics questionnaire and all subtests listed below) are administered using Gorilla Experiment Builder (www.gorilla.sc; Anwyl‐Irvine et al. 2020).

TABLE 1.

BRAIN‐online subtests a .

Task Description
BRAIN‐online Behavioral Questionnaire (BBQ) Participant's parent or caregiver rates the participant on a list of 14 behaviors commonly observed in individuals with FASD (e.g., hyperactivity, acting young for age).
Tapping Participant taps the spacebar with their index finger as fast as possible for 10s. Trials are repeated five times for each hand.
Reacting Participant taps the spacebar in response to a centrally‐presented stimulus as quickly as possible.
Inhibiting Participants respond to stimuli presented in a 2 × 2 grid. The task is to respond only when the stimulus is facing to the right and not when it is facing to the left.
Remembering Participant is shown a 4 × 4 grid of 14 hidden playing cards and two blank spaces. In the exposure trial, each card is revealed consecutively. This is followed by three learning trials, an immediate recall trial, and a delayed recall trial after approximately 10–15 min. In all trials, participants are asked where they saw each of the cards, and responses are made by clicking the box on the grid corresponding to the location of the card. Feedback is provided on learning trials, and the correct location of the card is shown. In the recall trials, no feedback is given.
Fishing Participant responds to a central stimulus that is presented with or without flanking stimuli. Flanking stimuli are either consistent or inconsistent with the central stimulus. The task is to indicate the direction of the central stimulus.
Stepping Stones Participant repeats a spatial span pattern that increases from 2 to 10 stimuli.
Connecting Participant connects circles in sequence in three trials: number order, letter order, or switching between numbers and letters in order.
Open in a new tab

Note: Table originally published in Mattson et al. (2026).

Abbreviations: BRAIN‐online, Brief Assessment of Individual Neurobehavior; BBQ, BRAIN‐online Behavioral Questionnaire.

a

Additional information about BRAIN‐online is available on request from the corresponding author.

1.2.2. In‐Person Neuropsychological Testing

As part of their participation in the CIFASD research project, participants completed a standardized, in‐person neuropsychological assessment that takes 2–4 h to complete. From this battery, measures were chosen for analysis that paralleled each BRAIN‐online subtest. The following subtests were chosen from standardized in‐person neuropsychological assessment tests: the Behavior Assessment Scale for Children, Third Edition (BASC‐3) Parent Rating Scales (Reynolds and Kamphaus 2015); Grooved Pegboard (Reitan and Wolfson 1985); NIH Toolbox Flanker Inhibitory Control and Attention Test (Gershon et al. 2013; Weintraub et al. 2013); NEPSY‐II Memory for Designs subtest (Korkman et al. 2007); California Verbal Learning Test, Children's Version (CVLT‐C; Delis et al. 2000); Wechsler Intelligence Scale for Children, Fifth Edition (WISC‐V) Spatial Span subtest (Wechsler 2014); Delis‐Kaplan Executive Functioning Scales (D‐KEFS) Trail‐Making subtests (Delis et al. 2001). See Table 2, for direct BRAIN‐online to in‐person assessment comparisons.

TABLE 2.

Corresponding BRAIN‐online and in‐person neuropsychological measures.

BRAIN‐online subtest In‐person neuropsychological subtests Domain
BBQ BASC‐3 Behavioral Symptoms Index Score Behavior
Tapping Grooved Pegboard Fine motor
Reacting NIH Toolbox Flanker Inhibitory Control and Attention Test Reaction time
Inhibiting NIH Toolbox Flanker Inhibitory Control and Attention Test Impulsivity
Remembering NEPSY‐II Memory for Designs; CVLT‐C Memory
Fishing NIH Toolbox Flanker Inhibitory Control and Attention Test Impulsivity
Stepping Stones WISC‐V Spatial Span Visual working memory
Connecting D‐KEFS Trail Making Executive functioning
Open in a new tab

Abbreviations: BASC‐3, Behavior Assessment Scale for Children, Third Edition; CVLT‐C, California Verbal Learning Test, Children's Version; D‐KEFS, Delis‐Kaplan Executive Functioning Scales; WISC‐V, Wechsler Intelligence Scale for Children, Fifth Edition.

1.3. Procedure

Caregivers provided consent for youth to participate in the larger CIFASD study, and youth participants age 7 and older provided assent. Caregivers completed measures of demographic information including PAE history and assessed youth behavioral challenges via online questionnaire. Youth completed in‐person neuropsychological testing in a laboratory setting and BRAIN‐online cognitive tasks at home. BRAIN‐online was administered using Gorilla Experiment Builder and completed either using a computer or a tablet with a connected keyboard. A link for BRAIN‐online was sent to caregivers of participants. Caregivers first completed the brief demographic questionnaire and the BRAIN‐online Behavior Questionnaire (BBQ), then the seven subtests were completed by the participant (i.e., child or adolescent) with supervision of their caregiver if needed. Caregivers were instructed not to help participants complete BRAIN‐online but were told they could aid in beginning the test, setting up the computer or tablet, reading the initial instructions, and prompting the participant to stay on task. Most participants completed BRAIN‐online before their in‐person neuropsychological assessment. Participants were seen at the Center for Behavior Teratology at San Diego State University (SDSU) for a 2–4‐h neuropsychological battery completed by a psychometrist or graduate student. Participants were paid for both the in‐person assessment and completion of BRAIN‐online.

1.4. Statistical Analyses

Analyses were run in SPSS v31.0.0. Group differences on demographic factors were tested using t‐tests and chi‐square tests of independence as appropriate. Because BRAIN‐online scores are not normed for demographic variables (e.g., age, sex), raw scores from the in‐person neuropsychological assessment were used to compare in‐person neuropsychological measures to BRAIN‐online subtests in correlation analyses. For direct comparisons on the equivalence and clinical utility of BRAIN‐online and in‐person neuropsychological assessment scores, z‐scores were calculated for all variables from both measures. These z‐scores were calculated based on only this sample and were not related to the standardization of any included measures. In addition, outlier scores that were 3.0 standard deviations (SDs) outside of the mean were excluded from analyses, which is consistent with current literature (Kline 2023). Due to varying age cutoffs for the in‐person neuropsychological measures, sample sizes varied by measure.

1.4.1. Correlations

The raw scores from all BRAIN‐online subtests were compared to raw scores from their parallel in‐person neuropsychological assessment measures using Pearson's correlations or chi‐square tests of independence if appropriate.

1.4.2. Equivalence Testing

To test whether there was measurement equivalence between in‐person neuropsychological assessment and BRAIN‐online subtests, equivalence testing was used (Lakens 2017). Difference scores were created by taking the absolute difference between analogous BRAIN‐online and in‐person neuropsychological test z‐scores, which were then compared with two one‐sided t‐tests (TOST; Lakens 2017).

1.4.3. Sensitivity, Specificity, and Accuracy Analyses

To analyze the clinical utility of BRAIN‐online in detecting neurobehavioral impairment, performance on each BRAIN‐online subtest was categorized as impaired or not impaired using criteria defined by Hoyme et al. (2016). Each score was converted to a z‐score based on whole group performance; if one or more measure was categorized as impaired (1.5 SD below group average per Hoyme et al.), neurobehavioral impairment was determined to be present. Using these criteria, classification (meeting criteria for neurobehavioral impairment vs. no neurobehavioral impairment) was compared to observed clinical diagnosis of FASD. Sensitivity (percentage of true positives identified [i.e., individuals identified as having neurobehavioral impairment by BRAIN‐online who also have a clinical FASD diagnosis]), specificity (percentage of true negatives identified [i.e., the number of individuals identified as not having neurobehavioral impairment by BRAIN‐online who do not have a clinical FASD diagnosis]), and accuracy (percentage of individuals correctly classified as FASD or no FASD based on whether or not they were classified as neurobehaviorally impaired) were calculated. This process was then repeated for in‐person neuropsychological assessment to compare sensitivity, specificity, and accuracy rates of both measures.

2. Results

2.1. Demographics

See Table 3 for participant demographics by group and t‐test/chi‐square results. In both groups, the average age of the participants was 9 years and approximately half of the participants were female. The groups did not significantly differ in age, sex, ethnicity, or family income. However, the CON group had a higher proportion of White participants than that of the AE group. IQ scores were derived from in‐person neuropsychological testing using the WISC‐V (Wechsler 2014). The AE group had a significantly lower average IQ score as compared to the CON group; however, the groups had a similar number of individuals with an IQ score at or below 77.5 (i.e., 1.5 SDs below average). Regarding other diagnoses, the AE group had a significantly higher percentage of individuals with an ADHD diagnosis or other neurodevelopmental diagnosis (e.g., ASD, epilepsy). Both groups completed BRAIN‐online under an h, with mean completion times of around 48 min. The groups did not significantly differ in the time between BRAIN‐online completion and in‐person neuropsychological assessment, which averaged from 62 to 69 days. The majority of participants in both groups completed BRAIN‐online prior to in‐person neuropsychological assessment; however, this was more common in the CON group than in the AE group.

TABLE 3.

Demographic data for participants with histories of prenatal alcohol exposure (the AE group) and controls (the CON group).

Variable AE (n = 101) CON (n = 56) Statistic p
Age at BRAIN‐online Completion [M (SD)] 9.5 (3.4) 9.1 (2.9) t = 0.786 0.433
Sex [n (%) Female] 53 (52.5%) 29 (51.8%) χ2 = 0.007 0.934
Race [n (%) White] 59 (58.4%) 43 (76.8%) χ2 = 5.341 0.021
Ethnicity [n (%) Hispanic/Latino] 44 (43.6%) 25 (44.6%) χ2 = 0.017 0.896
FASD Diagnosis a [n (%)] 86 (85.1%) 0 (0%) χ2 = 105.440 < 0.001
Average IQ [M (SD)] b 91.2 (13.8) 102.6 (16.2) t = −4.311 < 0.001
IQ < 77.5 [n (%)] b 18 (17.8%) 4 (7.1%) χ2 = 3.405 0.065
Handedness [n (%) Right] 83 (82.2%) 50 (89.3%) χ2 = 1.405 0.236
Family Income [n (%) ≤ $50,000] b 18 (17.8%) 9 (16.1%) χ2 = 0.111 0.739
ADHD Diagnosis c [n (%)] 56 (55.4%) 17 (30.4%) χ2 = 9.115 0.003
Other NDD Diagnosis c [n (%)] 75 (74.3%) 21 (37.5%) χ2 = 20.488 < 0.001
BRAIN‐online Completion Time [M (SD) min] 48.2 (14.1) 47.95 (16.6) t = 0.093 0.926
Difference in Years Between BRAIN‐online and Neuropsychological Test Completion [M (SD)] 0.19 (0.20) 0.17 (0.14) t = 0.568 0.571
BRAIN‐online Completion before Neuropsychological Test Completion [n (%)] 89 (88.1%) 56 (100%) χ2 = 7.204 0.007
Open in a new tab

Note: Bold values indicate p < 0.05.

Abbreviation: BRAIN‐online, Brief Assessment of Individual Neurobehavior, online version.

a

FASD diagnostic status was known for 87 participants in the AE group.

b

IQ missing for 8 participants; Family income missing for 2 participants.

c

Caregiver‐reported.

2.2. Correlations and Chi‐Square Tests

Correlations were calculated for each BRAIN‐online subtest and the parallel in‐person subtest (Table 4) and considered significant at p < 0.05. Sample sizes varied based on test completion, outlier removal, and age limitations of test administration. For behavior ratings, the number of items endorsed (out of 14) on the BBQ was correlated with the BASC‐3 Behavioral Symptoms Index raw score. To assess the clinical utility of the behavioral component of BRAIN‐online, the BBQ total score was dichotomized to reflect “elevated” behavior difficulties (four or more items endorsed) and “not elevated” behavior difficulties (three or fewer items endorsed) based on cutoff criteria set in a previous study (Mattson et al. 2026). This score was analyzed with the clinical cutoff (T ≥ 60) for the BASC‐3 as established by the test provider (Reynolds and Kamphaus 2015) using the chi‐square test of independence. A significant association was found, with an “elevated” classification on the BBQ associated with a clinical categorization on the BASC‐3 (ꭓ2(1) = 45.208, p < 0.001).

TABLE 4.

Correlations between BRAIN‐online and in‐person neuropsychological measures.

BRAIN‐online subtest In‐person neuropsychological subtest N Statistic p
BBQ Total Score BASC‐3 Behavioral Symptoms Index score (raw total) 153 χ2 = 45.208 < 0.001
Tapping—Dominant Hand Grooved Pegboard dominant hand raw score (total seconds) 129 r = 0.186 0.034
Tapping—Nondominant Hand Grooved Pegboard nondominant hand raw score (total seconds) 128 r = 0.077 0.388
Inhibiting—Total Correct on Incongruent Trials NIH Toolbox Flanker (% of correct responses on incongruent trials) 127 r = 0.074 0.411
Reacting—Median Reaction Time NIH Toolbox Flanker (median reaction time) 130 r = 0.521 < 0.001
Remembering—Learning 1–3 CVLT‐C Trials 1–5 Free Recall Total Correct Raw Score 146 r = 0.415 < 0.001
Remembering—Immediate NEPSY‐II Memory for Designs Immediate Recall 141 r = 0.545 < 0.001
CVLT‐C Short‐Delay Free Recall Total Correct Raw Score 145 r = 0.477 < 0.001
Remembering—Delay CVLT‐C Long Delay Free Recall Total Correct Raw Score 144 r = 0.452 < 0.001
NEPSY‐II Memory for Designs Delayed Recall 141 r = 0.523 < 0.001
Remembering—Retention CVLT‐C Retention = Long Delay Recall Score/Short‐Delay Recall Score 127 r = 0.008 0.929
Fishing—Total Accuracy NIH Toolbox Flanker (% of correct responses on all trials) 131 r = 0.35 < 0.001
Stepping Stones—Span Length WISC‐V Spatial Span LSSPf 125 r = 0.49 < 0.001
Stepping Stones—Trials Correct WISC‐V Spatial Span Forward Total Raw Score 127 r = 0.545 < 0.001
Connecting—Numbers D‐KEFS Trail Making Number Sequencing Raw Score 97 r = 0.407 < 0.001
Connecting—Letters D‐KEFS Trail Making Letter Sequencing Raw Score 99 r = 0.074 0.465
Connecting—Switching D‐KEFS Trail Making Number‐Letter Switching Raw Score 99 r = 0.378 < 0.001
Open in a new tab

Note: Bold values indicate p < 0.05.

Abbreviations: BRAIN‐online, Brief Assessment of Individual Neurobehavior, online version; BASC‐3, Behavior Assessment Scale for Children, Third Edition; CVLT‐C, California Verbal Learning Test, Children's Version; D‐KEFS, Delis‐Kaplan Executive Functioning Scales; WISC‐V, Wechsler Intelligence Scale for Children, Fifth Edition.

For the cognitive subtests, the following BRAIN‐online subtests were significantly correlated with their in‐person counterparts: Tapping (Dominant Hand), Reacting (Median Reaction Time), Remembering (Learning Trials, Immediate Recall, Delayed Recall), Fishing (Total Accuracy), Stepping Stones (Span Length, Trials Correct), Connecting (Numbers, Switching). These subtests were not significantly correlated with their in‐person counterparts: Tapping (Nondominant Hand), Inhibiting (Total Correct Responses on Incongruent Trials), Remembering (Retention), Connecting (Letters). Retention scores were uniformly high for both measures, as in previous studies (Crocker et al. 2011; Roebuck‐Spencer and Mattson 2004).

2.3. Equivalence Testing

Measurement equivalence was tested using equivalence testing (Lakens 2017). This procedure was used to determine whether the effect size of the difference between the BRAIN‐online and in‐person measures was negligible and therefore, both measures assess the same domains of cognitive functioning. If the difference between BRAIN‐online scores and in‐person neuropsychological assessment measure scores falls below the Smallest Effect Size of Interest (SESOI), which is determined by the literature and power analyses, then these measures can be considered to be equivalent (Lakens 2017).

Because unstandardized scores were used for each of the measures and scales differ for each measure, z‐scores based on whole group performance were used. These z‐scores were tested using the TOST procedure, which involves assessing the mean difference of the z‐scores using two one‐sided t‐tests (Lakens 2017). In order to compare if the effect size of the difference of the two measures was negligible (i.e., the tests were equivalent), the SESOI was calculated using G*Power (Faul et al. 2007) based on the current sample size, power of 0.80, and an alpha of 0.05. The SESOI was thus determined to be 0.25, which represents a small effect size based on Cohen's d (Cohen 1988). Therefore, the upper and lower bounds tested in the TOST procedure were 0.25 and −0.25, respectively. If both the t‐score for the lower and upper bounds are statistically significant and in the predicted directions (i.e., t‐score for the upper bound is less than 0 and t‐score for the lower bound is above 0), the measures are deemed to be equivalent (Lakens 2017). t‐tests were assessed using an alpha of 0.05 (90% Confidence Interval [CI]). Results of both one‐sided t‐tests can be found in Table 5.

TABLE 5.

TOST results for equivalence testing for BRAIN‐online and in‐person neuropsychological test performance.

BRAIN‐online subtest In‐person neuropsychological subtest N t lower (p) t upper (p) Equivalence
BBQ Total Hits (Yes's) BASC‐3 externalizing score (raw total) 153 4.541 (< 0.001) −3.991 (< 0.001) Equivalent
Tapping—Dominant Hand Grooved Pegboard dominant hand raw score (total seconds) 129 2.242 (0.013) −3.269 (< 0.001) Equivalent
Tapping—Nondominant Hand Grooved Pegboard nondominant hand raw score (total seconds) 128 2.315 (0.011) −3.351 (< 0.001) Equivalent
Inhibiting—Total Correct Hits on Right NIH Toolbox Flanker Subtest (% of total correct hits on incongruent trials) 127 3.627 (< 0.001) −3.052 (0.001) Equivalent
Reacting—Median Reaction Time NIH Toolbox Flanker Subtest Median Reaction Time 129 3.656 (< 0.001) −2.928 (0.002) Equivalent
Remembering—Learning 1–3 CVLT‐C Trials 1–5 Free Recall Total Correct Raw Score 146 2.325 (0.011) −3.256 (< 0.001) Equivalent
Remembering—Immediate CVLT‐C Short‐Delay Free Recall Total Correct Raw Score 145 3.020 (0.001) −2.909 (0.002) Equivalent
Remembering—Delay CVLT‐C Long Delay Free Recall Total Correct Raw Score 144 2.815 (0.003) −2.974 (0.002) Equivalent
Remembering—Retention CVLT‐C Retention = Long Delay Recall Score/Trial 5 Recall Score 127 1.809 (0.036) −3.095 (0.001) Equivalent
Fishing—Total Accuracy NIH Toolbox Flanker Total Raw Score 131 3.153 (0.001) −2.784 (0.003) Equivalent
Stepping Stones—Span Length WISC‐V Spatial Span LSSPf 125 4.106 (< 0.001) −2.021 (0.023) Equivalent
Stepping Stones—Trials Correct WISC‐V Spatial Span Forward Total Raw Score 127 4.308 (< 0.001) −1.755 (0.041) Equivalent
Connecting—Numbers D‐KEFS Trail Making Number Sequencing Raw Score 97 1.164 (0.124) −5.828 (< 0.001) Not Equivalent
Connecting—Letters D‐KEFS Trail Making Letter Sequencing Raw Score 99 −0.613 (0.271) −6.045 (< 0.001) Not Equivalent
Connecting—Switching D‐KEFS Trail Making Number‐Letter Switching Raw Score 99 −1.740 (0.043) −7.106 (< 0.001) Not Equivalent
Open in a new tab

Note: SESOI = ±0.25. Equivalence is supported when both t lower and t upper are significant at α = 0.05 (90% CI).

Abbreviations: BASC‐3, Behavior Assessment Scale for Children, Third Edition; BRAIN‐online, Brief Assessment of Individual Neurobehavior, online version; CVLT‐C, California Verbal Learning Test, Children's Version; D‐KEFS, Delis‐Kaplan Executive Functioning Scales; TOST, Two One‐Sided t‐tests; WISC‐V, Wechsler Intelligence Scale for Children, Fifth Edition.

Measurement equivalence was found for seven out of eight of the BRAIN‐online subtests and their corresponding in‐person neuropsychological measures. Equivalence was found for the following pairs of tests: BBQ and BASC‐3 (behavioral symptoms); Tapping and Grooved Pegboard, both dominant and nondominant hands (fine motor speed); Inhibiting and NIH Toolbox Flanker percent correct responses (inhibitory control); Reacting and NIH Toolbox Flanker response time (reaction time); all of the Remembering and CVLT‐C scores (learning, immediate & delayed memory, retention); Stepping Stones and WISC‐V Spatial Span (visual working memory).

Equivalence was not found for Connecting and D‐KEFS Trail‐Making (processing speed and executive functioning). Average z‐scores were higher for the Numbers Connecting subtest (M = −0.061) when compared to the D‐KEFS Number sequencing subtest (M = −0.115), reflecting that participants had a higher average processing speed for this specific task. In contrast, the average z‐score from Letter sequencing (M = −0.128) trial from the BRAIN‐online Connecting subtest was lower than the average z‐score from the Letter sequencing (M = −0.093) subtest from the D‐KEFS and the Switching sequencing (M = −0.105) from the BRAIN‐online Connecting subtest was lower than the average z‐score from the Number‐Letter Switching sequencing (M = 0.000) subtest from the D‐KEFS, indicating that BRAIN‐online captured lower processing speed and decreased executive functioning on these specific tasks. While in some cases, the mean differences are small, the 90% CI falls outside the SESOI (i.e., −0.25 to 0.25), indicating nonequivalence. Similarly, though the TOST results for the Switching Connecting subtest and D‐KEFS Number‐Letter Switching task are both significant, the 90% CI falls outside of the SESOI, suggesting that these two tasks cannot be considered to be equivalent. All TOST results can be found in Table 5 and Figure 1.

FIGURE 1.

FIGURE 1

Open in a new tab

Mean Differences of z‐scores of BRAIN‐online and In‐Person Neuropsychological Measures. Difference scores were calculated with BRAIN‐online and in‐person neuropsychological z‐scores for each participant, and then these scores were averaged and tested using the Two One‐Sided Tests (TOST) Procedure. If the mean difference scores and their 90% confidence intervals (CIs) fell within the Smallest Effect Size of Interest (SESOI), which was identified as −0.25 to 0.25, the two measures were determined to be equivalent.

2.4. Sensitivity, Specificity, and Accuracy Analyses

To assess the clinical utility of BRAIN‐online, BRAIN‐online scores were used to determine neurobehavioral impairment as defined by the Hoyme FASD criteria (Hoyme et al. 2016). According to these criteria, individuals must either show evidence of global impairment or deficit in at least one neurobehavioral domain (including executive functioning, learning, memory, or visual–spatial) at or below 1.5 SD from the mean or evidence of behavioral challenge in at least one domain of self‐regulation (including mood or behavioral regulation, attention, or impulse control) at or below 1.5 SD from the mean (Hoyme et al. 2016). Therefore, if individuals had z‐scores at or below 1.5 SD from the mean on any of the BRAIN‐online subtests (including the BBQ), this criterion was met, and they were classified as “at‐risk” for FASD. Then, these classification results were compared to clinical diagnosis of FASD (i.e., FASD vs. not FASD) in our sensitivity, specificity, and accuracy analyses. The FASD group included 85% (n = 86) of the AE group that met criteria for an FASD; the not FASD group included controls and participants in the AE group that did not meet criteria for an FASD (n = 1) or did not have enough information for a determination of FASD (n = 14). Sensitivity was calculated by dividing the number of true positives (i.e., individuals in the FASD group identified as at‐risk on BRAIN‐online) by the total number of positives (i.e., all individuals identified as at‐risk). Specificity was calculated by dividing the number of true negatives (i.e., individuals in the no FASD group identified as not at‐risk on BRAIN‐online) by the total number of negatives (i.e., all individuals identified as not at‐risk). Accuracy was calculated by dividing the total number of true negatives and true positives by the total sample. This procedure was repeated for in‐person neuropsychological measures to compare sensitivity, specificity, and accuracy results from the two approaches.

See Table 6 for contingency tables and Table 7 for the corresponding sensitivity, specificity, and accuracy results. Results showed that both measures had comparable though low accuracy rates. Although both BRAIN‐online and in‐person neuropsychological measures had high sensitivity rates, BRAIN‐online had a higher sensitivity value, detecting an increased number of individuals with FASD. Both BRAIN‐online and in‐person neuropsychological assessment had poor specificity rates, with in‐person neuropsychological assessment having a slightly higher specificity value, indicating that both measures may capture individuals with behavioral or cognitive difficulties that do not necessarily have FASD. Further analyses were done excluding the individuals with PAE that did not meet criteria for FASD (n = 15); however, results of the sensitivity, specificity, and accuracy analyses were minimally affected.

TABLE 6.

Contingency tables for (a) in‐person neuropsychological assessment and (b) BRAIN‐online using the 1.5 SD cutoff for impairment based on Hoyme FASD criteria (Hoyme et al. 2016) to compare to FASD diagnoses.

FASD No FASD
a. In‐person neuropsychological assessment
Meets criteria for neurobehavioral impairment based on in‐person neuropsychological assessment 72 47
Does not meet criteria for neurobehavioral impairment based on in‐person neuropsychological assessment 14 24
Total 86 71
b. BRAIN‐online
Meets criteria for neurobehavioral impairment based on BRAIN‐online 79 53
Does not meet criteria for neurobehavioral impairment based on BRAIN‐online 7 18
Total 86 71
Open in a new tab

Abbreviation: BRAIN‐online, Brief Assessment of Individual Neurobehavior—online version.

TABLE 7.

Results of sensitivity, specificity, and accuracy analyses using the 1.5 SD cutoff for impairment based on Hoyme FASD criteria (Hoyme et al. 2016) to compare to clinical FASD diagnoses.

Measure In‐person assessment BRAIN‐online
Sensitivity 83.72% 91.86%
Specificity 33.80% 25.35%
Accuracy 61.15% 61.78%
Positive Predictive Values (PPV) 60.50% 59.85%
Negative Predictive Values (NPV) 63.16% 72.00%
Open in a new tab

Abbreviations: NPV, Negative Predictive Value; PPV, Positive Predictive Value.

3. Discussion

This study examined the validity and accuracy of an online FASD screening tool, BRAIN‐online, to identify children and adolescents with and without PAE in comparison with in‐person neuropsychological assessment, the gold‐standard method for FASD diagnosis. Because diagnosis of an FASD requires a physical examination and comprehensive neuropsychological testing to evaluate cognitive and behavioral symptoms, this process requires considerable time and monetary costs that place a significant burden on both children and their families (Kent et al. 2023). In addition, the costs of FASD health care, including diagnosis, pose a significant economic burden internationally (Greenmyer et al. 2018), some of which may be alleviated with more efficient and accessible screening and diagnostic tools. BRAIN‐online was created to screen for cognitive and behavioral symptoms known to be associated with FASD with the goal of potentially streamlining the diagnostic process for individuals who have deficits and symptoms associated with FASD. In doing so, BRAIN‐online can potentially reduce clinic wait times by screening out individuals who do not need further FASD assessment and shorten assessment appointments by preliminarily highlighting cognitive and behavioral difficulties to which the clinician should attend. This tool has been found to adequately differentiate individuals with PAE from those without (Mattson et al. 2026), but no previous study has compared BRAIN‐online to the gold‐standard, in‐person neuropsychological assessment measures. To evaluate criterion validity and the ability of BRAIN‐online to assess domains of cognitive and behavioral functioning known to be associated with FASD, scores from the seven BRAIN‐online subtests and behavioral questionnaire were compared to parallel scores from in‐person neuropsychological assessment measures and one behavioral questionnaire.

Of the eight subtests of BRAIN‐online, seven were significantly correlated with in‐person measures, indicating that BRAIN‐online subtests capture similar individual performance to in‐person neuropsychological assessment. These results suggest that BRAIN‐online has the ability to capture functioning in cognitive and behavioral domains associated with FASD with cost‐ and time‐effective assessments that can be completed at home in under an hour. In addition, results of the equivalence testing showed seven of the eight BRAIN‐online subtests (including the behavioral questionnaire) could be deemed equivalent to their in‐person counterparts. Meeting equivalence with in‐person neuropsychological measures suggests that BRAIN‐online tasks are able to assess similar constructs as gold‐standard methods and, in turn, are validated to identify the need for further FASD assessment.

Importantly, some of the BRAIN‐online subtests were not found to be significantly correlated with their parallel in‐person testing. Nondominant hand performance on the Tapping subtest, measuring fine motor speed, was not significantly correlated with nondominant hand performance on the Grooved Pegboard task, although they were found to be equivalent. The lack of significant correlation may relate to the methodological differences in these two tasks in measuring fine motor speed; Tapping assesses pure motor speed while Grooved Pegboard involves visual–spatial skills and relies more heavily on dexterity. However, importantly, the dominant hand performance on both measures was correlated and equivalent, indicating potentially less variability in dominant hand performance. The Inhibiting task, which measured inhibitory control, was not correlated with scores from the NIH Toolbox Flanker subtest, although they were found to be equivalent. The lack of significant correlation may be due to variability from different testing environments that influence one's regulatory responses. Another reason for this difference may be that in this task, the score used to capture inhibitory control was the number of correct responses when stimuli were facing left. In order to receive a correct score on these trials, participants had to inhibit their response and not touch the screen. In contrast, inhibitory control as captured by the NIH Toolbox Flanker subtest was captured through participant response to incongruent stimuli. Therefore, the lack of correlation of these two subtests may be due to the difference in methods of assessing inhibitory control. Measures of retention on the Remembering subtest and CVLT‐C were also not significantly correlated although they were found to be equivalent. It is possible that the lack of correlation between these two measures was due to a restriction in range of scores as retention was high for both measures. However, despite these two scores not being correlated, the other Remembering and CVLT‐C scores were correlated, indicating that the Remembering task scores are still able to capture other facets of memory functioning. In addition, previous analyses of the BRAIN‐online Retention score did not result in significant differences between individuals with PAE and individuals without (Mattson et al. 2026), and similar results have been seen with retention using the CVLT‐C (Crocker et al. 2011; Roebuck‐Spencer and Mattson 2004). Therefore, memory retention may not be a meaningful measure in identifying individuals who may have an FASD. The Letter Sequencing task from BRAIN‐online was not correlated with the Letter sequencing task from the D‐KEFS. While this lack of correlation may be due to differences in test mechanism (i.e., paper and pencil vs. electronic device), the other two connecting tasks were significantly correlated with their D‐KEFS counterparts. Overall, while certain subtests may need better parallel assessment measures for comparison, significant correlations in the majority of the BRAIN‐online and in‐person neuropsychological tasks is a strong indicator of similar individual performances across the two measures.

Regarding equivalence, the Connecting subtests did not meet equivalence with their parallel in‐person neuropsychological measures, although they were found to be significantly correlated. While some of these measures produce similar scores to their in‐person measure counterparts on an individual level, there are average differences of scores between the two that are too large to be labeled as equivalent to in‐person assessments. There are many reasons for this finding, such as differences in the environment (home vs. testing room) and differences in how the test is conducted (computer vs. paper‐pencil). For example, in the processing speed domain, individuals on average may have a different completion time on Connecting, a computerized task, as compared to D‐KEFS Trail‐Making, in which individuals have to manually draw connections. This was reflected in their average z‐scores, with the Numbers task of the BRAIN‐online Connecting subtest capturing faster processing speed than their counterparts in the D‐KEFS. Importantly, however, though the Connecting and D‐KEFS Trail‐Making tasks may be measuring completion time differently on average, the scores from these measures are still correlated, which means that Connecting scores are able to capture underlying processing speed and executive functioning and therefore can indicate the need for further FASD assessment. The goal of FASD screening using BRAIN‐online is not to replace in‐person neuropsychological assessment, and therefore, correlation of different subtests may be more informative for assessment of this screening tool.

Interestingly, there were a few discrepancies between our correlational analyses and equivalence testing. For example, though scores on the Inhibiting task were not correlated with NIH Toolbox Flanker scores of correct responses on incongruent trials, these scores were found to be equivalent. While correlations indicate whether there is an association between the two scores on an individual level, equivalence testing indicates how closely two different measures can capture scores in a domain of functioning in the sample on average. In the case of inhibition, results suggest that the Inhibiting and NIH Toolbox Flanker scores were not similar per person, but overall the two measures assess inhibitory control closely enough (smaller than a small effect size) to be considered equivalent. This may be due to within‐person variability that can be influenced by factors such as the testing environment. Importantly, though it is helpful to see that certain BRAIN‐online subtests meet equivalence with gold‐standard measures, the goal of BRAIN‐online is to screen for PAE‐related symptoms, not to replace comprehensive neuropsychological assessment.

Finally, we assessed the clinical utility of BRAIN‐online by using cutoff scores specified by Hoyme et al.'s (2016) criteria, specifically evaluating whether individuals in each group were accurately classified as having a higher likelihood for FASD using the neurobehavioral impairment component of the FASD diagnosis. Although diagnosis of an FASD requires cognitive and/or behavioral assessment, the neuropsychological assessment measures used in this study did not include variables commonly used for FASD diagnosis (e.g., IQ score) and therefore allowed us to assess how these specific domains mirrored in BRAIN‐online could predict FASD classification. Results of these analyses showed that both BRAIN‐online and in‐person neuropsychological assessment had low classification accuracy but high sensitivity rates when compared to clinical FASD diagnoses. This finding not only supports BRAIN‐online as an effective screening tool for FASD‐related challenges but also shows that BRAIN‐online can identify neurobehavioral impairment comparably to gold‐standard measures in the context of PAE‐related symptoms, though both BRAIN‐online and in‐person measures should be further investigated to increase accuracy in identifying FASD. It is important to note that the assessment of cognitive and behavioral symptoms through neuropsychological assessment is one part of a multidisciplinary process that, when combined, may result in higher diagnostic accuracy. Results of the specificity analyses revealed poor specificity for both BRAIN‐online and in‐person neuropsychological assessment, with in‐person neuropsychological assessment having slightly higher specificity. This may be due to the heterogeneity of our control group, which did not exclude individuals with other neurodevelopmental disorders and is a strength of our study design, leading to an increase in individuals falsely identified as FASD‐risk by both measures. While it is important to address comorbidity of symptoms for diagnostics and intervention, the goal of BRAIN‐online is to screen individuals in need of further assessment for FASD‐related difficulties and therefore is more beneficial to prioritize sensitivity over specificity for this tool. Other FASD‐related factors, such as facial dysmorphology, will be considered upon further assessment to help differentiate individuals with similar needs. Importantly, when considering BRAIN‐online as a whole in the original study (Mattson et al. 2026), specificity was considerably higher (89.6%) than in the current analysis (25.4%), suggesting that further study is needed to determine if study‐specific factors contributed to this large difference in specificity rates. For example, the studies differed in sample size and how impairment was determined. Perhaps using a single indicator of impairment (1.5 SD below the mean on one domain) is too liberal to determine clinical significance. Importantly for the current study, the specificity for in‐person neuropsychological assessment was also low, possibly due to the use of raw scores and whole group performance as a reference for impairment that may categorize too many individuals as “impaired.” In summary, while both measures yielded low specificity rates which are different from previous studies, low specificity does not diminish the ability of BRAIN‐online as a screening tool where a more liberal approach to classification is appropriate.

Although this study has numerous strengths including a well‐characterized sample with both prenatally exposed and control groups, there are important limitations. First, we did not have adequate direct comparisons for all BRAIN‐online subtests, such as the Inhibiting subtest. Although a score was derived from the NIH Toolbox Flanker subtest that measured inhibitory control, it did not measure inhibitory control in the same way as the Inhibiting task, which required participants to inhibit their responses during certain trials. In addition, sample sizes varied due to age cutoffs for different measures. For example, though our study included children in the range of 5–17 years, only children 8 years and above could complete the D‐KEFS measures, resulting in a smaller sample size. Although even the smallest sample size was powered for our analyses, having a larger sample size may help us better understand the relationship between the online and in‐person measures. In addition, raw scores were used for all of our analyses, and thus, our analyses did not account for demographic variables such as age and gender/sex. With a larger sample size, creating standardized scores for BRAIN‐online may be helpful in identifying areas of challenges that can be comparable to other standardized measures. Using raw scores to assess impairment may have interacted with our classification results, and therefore, assessing standardized neuropsychological assessment scores for classification accuracy may give a better picture of how cognitive and behavioral measures may affect specificity. Further, impairment was defined using the entire sample, including both clinical and nonclinical youth. This has the potential to skew the comparison for impairment as individuals with cognitive and behavioral difficulties were also used to define the comparison sample. In the future, a larger sample size could be used to create a comparable nonclinical normative sample to assess whether youth with PAE meet the “impairment” threshold defined by diagnostic criteria. Unequal group size may also influence results for the classification analyses and should therefore be considered in future analyses (Zahn 2010). This analysis excluded data that was ±3.0 SDs, which may affect our ability to assess the clinical utility of BRAIN‐online as extreme scores may be clinically meaningful. However, because BRAIN‐online is an unstandardized measure, data standardization techniques that are consistent with the current literature (Kline 2023) were used. Due to the online nature of BRAIN‐online that requires internet access and an electronic device (i.e., computer or tablet), factors such as internet connectivity issues and lack of access of electronic devices may act as a barrier to completing BRAIN‐online (Walker et al. 2023). This is an important consideration especially as children with FASD may have access to fewer resources (Sood et al. 2001). Although not all of these barriers are avoidable, BRAIN‐online increases accessibility through online administration that can be accessed from any computer or tablet in a space that provides internet access (e.g., public library).

There are many future directions for a validated screening measure for FASD. While previous studies have evaluated whether BRAIN‐online can identify challenges associated with PAE, questions related to how well BRAIN‐online can differentiate individuals with symptoms associated with FASD from other neurodevelopmental disorders remain, which are common in individuals with FASD (Weyrauch et al. 2017). Having a screening measure specific to FASD can better help identify individuals with FASD‐specific symptoms and streamline diagnostic processes, reducing clinic wait times for FASD‐specific assessments. Future research should also assess whether BRAIN‐online results map onto neurodevelopmental processes via neural imaging or functional connectivity studies. In doing so, BRAIN‐online scores may be able to represent similar neurodevelopmental processes experienced by individuals with FASD. This connection would allow for increased precision in FASD diagnosis, as functional magnetic resonance imaging (fMRI) has been used to find dose‐dependent relationships on network connectivity (Wozniak et al. 2019). Reflecting these differences in connectivity through BRAIN‐online will further validate this measure as an FASD screening tool. Finally, the clinical utility of BRAIN‐online should be studied in the context of diagnostic criteria in different countries, as there are national differences in impairment cutoffs and guidelines that are reflective of the differing national drinking patterns. By assessing BRAIN‐online in differing diagnostic contexts, BRAIN‐online can be validated as a universal screening tool for FASD.

In conclusion, BRAIN‐online is a brief, low‐cost online tool that can make screening for FASD more accessible and efficient. Use of BRAIN‐online can streamline diagnostic assessment for some and suggest that alternative diagnoses should be considered for others. Our results indicate that performance on BRAIN‐online is correlated with in‐person neuropsychological assessment and results demonstrate measurement equivalence on most measures. Use of this online resource may lead to increased diagnostic accuracy and efficiency for individuals with PAE, potentially leading to FASD‐specific intervention and ultimately, lifelong well‐being for individuals with FASD.

Funding

All or part of this work was done in conjunction with the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD, https://doi.org/10.5967/ntw9‐h991), which is funded by grants from the National Institute on Alcohol Abuse and Alcoholism (NIAAA). Research described in this paper was supported by NIAAA grant number U01 AA014834. Additional support was provided by NIAAA grants U24AA014811, U01AA026102, U01AA030164, F31AA031895, and T32AA013525. The NIH had no role in the design and conduct of the study.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

All or part of this work was done in conjunction with the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD, https://doi.org/10.5967/ntw9‐h991), which is funded by grants from the National Institute on Alcohol Abuse and Alcoholism (NIAAA). Additional information about CIFASD, including about data sharing opportunities, can be found at https://www.cifasd.org. Data from this study are part of Phase 4 (2017‐2022) and Phase 5 (2022‐2027).

References

  1. Alfano, A. R. , Concepcion I., Espinosa A., and Menendez F.. 2024. “Pediatric Language Assessments via Telehealth: A Systematic Review.” Journal of Telemedicine and Telecare 30: 1065–1076. [DOI] [PubMed] [Google Scholar]
  2. American Psychiatric Association . 2022. Diagnostic and Statistical Manual of Mental Disorders. 5th ed., text rev. ed. American Psychiatric Association Publishing. [Google Scholar]
  3. Anwyl‐Irvine, A. L. , Massonnie J., Flitton A., Kirkham N., and Evershed J. K.. 2020. “Gorilla in Our Midst: An Online Behavioral Experiment Builder.” Behavior Research Methods 52: 388–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Astley, S. J. , Stachowiak J., Clarren S. K., and Clausen C.. 2002. “Application of the Fetal Alcohol Syndrome Facial Photographic Screening Tool in a Foster Care Population.” Journal of Pediatrics 141: 712–717. [DOI] [PubMed] [Google Scholar]
  5. Brown, T. , and Zakzanis K. K.. 2025. “A Review of the Reliability of Remote Neuropsychological Assessment.” Applied Neuropsychology. Adult 32: 1536–1542. [DOI] [PubMed] [Google Scholar]
  6. Chasnoff, I. J. , Wells A. M., and King L.. 2015. “Misdiagnosis and Missed Diagnoses in Foster and Adopted Children With Prenatal Alcohol Exposure.” Pediatrics 135: 264–270. [DOI] [PubMed] [Google Scholar]
  7. Cohen, J. 1988. Statistical Power Analysis for the Behavioral Sciences. Lawrence Erlbaum Associates. [Google Scholar]
  8. Coles, C. D. , Bandoli G., Kable J. A., Del Campo M., Suttie M., and Chambers C. D.. 2023. “Comparison of Three Systems for the Diagnosis of Fetal Alcohol Spectrum Disorders in a Community Sample.” Alcoholism, Clinical and Experimental Research 47: 370–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Crocker, N. , Vaurio L., Riley E. P., and Mattson S. N.. 2011. “Comparison of Verbal Learning and Memory in Children With Heavy Prenatal Alcohol Exposure or Attention‐Deficit/Hyperactivity Disorder.” Alcoholism, Clinical and Experimental Research 35: 1114–1121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Delis, D. , Kramer J. H., Kaplan E., and Ober B. A.. 2000. California Verbal Learning Test—Second Edition (CVLT‐II) Manual.
  11. Delis, D. C. , Kaplan E., and Kramer J. H.. 2001. The Delis‐Kaplan Executive Function System: Examiner's Manual.
  12. Engesether, B. , Hoffner M., Johnson E., Klug M. G., Popova S., and Burd L.. 2024. “Prevalence of Fetal Alcohol Spectrum Disorder in Foster Care: A Scoping Review.” Alcoholism, Clinical and Experimental Research 48: 1443–1450. [DOI] [PubMed] [Google Scholar]
  13. Faul, F. , Erdfelder E., Lang A. G., and Buchner A.. 2007. “G*Power 3: A Flexible Statistical Power Analysis Program for the Social, Behavioral, and Biomedical Sciences.” Behavior Research Methods 39: 175–191. [DOI] [PubMed] [Google Scholar]
  14. Gershon, R. C. , Wagster M. V., Hendrie H. C., Fox N. A., Cook K. F., and Nowinski C. J.. 2013. “NIH Toolbox for Assessment of Neurological and Behavioral Function.” Neurology 80: S2–S6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Greenmyer, J. R. , Klug M. G., Kambeitz C., Popova S., and Burd L.. 2018. “A Multicountry Updated Assessment of the Economic Impact of Fetal Alcohol Spectrum Disorder: Costs for Children and Adults.” Journal of Addiction Medicine 12: 466–473. [DOI] [PubMed] [Google Scholar]
  16. Hoyme, H. E. , Kalberg W. O., Elliott A. J., et al. 2016. “Updated Clinical Guidelines for Diagnosing Fetal Alcohol Spectrum Disorders.” Pediatrics 138: e20154256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jones, K. L. , Robinson L. K., Bakhireva L. N., et al. 2006. “Accuracy of the Diagnosis of Physical Features of Fetal Alcohol Syndrome by Pediatricians After Specialized Training.” Pediatrics 118: e1734–e1738. [DOI] [PubMed] [Google Scholar]
  18. Kautz‐Turnbull, C. , Rockhold M., Handley E. D., Olson H. C., and Petrenko C.. 2023. “Adverse Childhood Experiences in Children With Fetal Alcohol Spectrum Disorders and Their Effects on Behavior.” Alcoholism, Clinical and Experimental Research 47: 577–588. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kent, N. , Hayes N., Young S., et al. 2023. “Exploring Resource Implications and Models of Care for Assessment and Diagnosis of Fetal Alcohol Spectrum Disorder: A Scoping Review.” Alcoholism, Clinical and Experimental Research 47: 2022–2032. [DOI] [PubMed] [Google Scholar]
  20. Kline, R. B. 2023. Principles and Practice of Structural Equation Modeling. Guilford Press. [Google Scholar]
  21. Korkman, M. , Kirk U., and Kemp S.. 2007. NEPSY‐Ii. Administration Manual. 2nd ed.
  22. Kurniadi, N. E. , and Davis J. J.. 2022. “Comparison Shopping: Is Neuropsychological Evaluation More Expensive Than Neuroimaging?” Clinical Neuropsychologist 36: 2061–2072. [DOI] [PubMed] [Google Scholar]
  23. Lakens, D. 2017. “Equivalence Tests: A Practical Primer for.” Social Psychological and Personality Science 8: 355–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Maenner, M. J. , Shaw K. A., Baio J., et al. 2020. “Prevalence of Autism Spectrum Disorder Among Children Aged 8 Years—Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2016.” MMWR Surveillance Summaries 69: 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Mattson, S. N. , Bernes G. A., and Doyle L. R.. 2019. “Fetal Alcohol Spectrum Disorders: A Review of the Neurobehavioral Deficits Associated With Prenatal Alcohol Exposure.” Alcoholism, Clinical and Experimental Research 43: 1046–1062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Mattson, S. N. , Foroud T., Sowell E. R., et al. 2010. “Collaborative Initiative on Fetal Alcohol Spectrum Disorders: Methodology of Clinical Projects.” Alcohol 44: 635–641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mattson, S. N. , Hyland M. T., Chockalingam G., et al. 2026. “BRAIN‐Online: An Online Cognitive and Behavioral Screening Tool for Fetal Alcohol Spectrum Disorders.” Alcoholism, Clinical and Experimental Research 50: e70261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. May, P. A. , Chambers C. D., Kalberg W. O., et al. 2018. “Prevalence of Fetal Alcohol Spectrum Disorders in 4 US Communities.” JAMA 319: 474–482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. McNally Keehn, R. , Enneking B., Ryan T., et al. 2023. “Tele‐Assessment of Young Children Referred for Autism Spectrum Disorder Evaluation During COVID‐19: Associations Among Clinical Characteristics and Diagnostic Outcome.” Autism 27: 1362–1376. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Moore, E. M. , and Riley E. P.. 2015. “What Happens When Children With Fetal Alcohol Spectrum Disorders Become Adults?” Current Developmental Disorders Reports 2: 219–227. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Popova, S. , Dozet D., Temple V., Mcfarlane A., Cook J., and Burd L.. 2024. “Fetal Alcohol Spectrum Disorder Diagnostic Clinic Capacity in Canadian Provinces and Territories.” PLoS One 19: e0301615. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Reitan, R. M. , and Wolfson D.. 1985. The Halstead‐Reitan Neuropsychological Test Battery: Theory and Clinical Interpretation. Neuropsychology Press. [Google Scholar]
  33. Reynolds, C. R. , and Kamphaus R. W.. 2015. Behavior Assessment System for Children (3rd Ed.).
  34. Ritfeld, G. J. , Kable J. A., Holton J. E., and Coles C. D.. 2022. “Psychopharmacological Treatments in Children With Fetal Alcohol Spectrum Disorders: A Review.” Child Psychiatry and Human Development 53: 268–277. [DOI] [PubMed] [Google Scholar]
  35. Ritfeld, G. J. , Kable J. A., Holton J. E., and Coles C. D.. 2024. “Effectiveness of Psychotropic Medications in Children With Prenatal Alcohol and Drug Exposures: A Case Series and Model of Care.” Child Psychiatry and Human Development 55: 744–753. [DOI] [PubMed] [Google Scholar]
  36. Ritter, C. , Mclachlan K., De Wolff O., Mcmurtry C. M., and Lumley M.. 2025. “A Content Analysis of Caregiver Perspectives on Strengths and Protective Factors in Youth With FASD and Criminal Legal System Involvement.” Health Justice 13: 76. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rockhold, M. N. , Gimbel B. A., Richardson A. A., et al. 2024. “Racial and Ethnic Disparities in Psychological Care for Individuals With FASD: A Dis/Ability Studies and Critical Race Theory Perspective Toward Improving Prevention, Assessment/Diagnosis, and Intervention.” Frontiers in Public Health 12: 1355802. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Roebuck‐Spencer, T. M. , and Mattson S. N.. 2004. “Implicit Strategy Affects Learning in Children With Heavy Prenatal Alcohol Exposure.” Alcoholism, Clinical and Experimental Research 28: 1424–1431. [DOI] [PubMed] [Google Scholar]
  39. Ros‐Demarize, R. , Chung P., and Stewart R.. 2021. “Pediatric Behavioral Telehealth in the Age of COVID‐19: Brief Evidence Review and Practice Considerations.” Current Problems in Pediatric and Adolescent Health Care 51: 100949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Schaefer, L. A. , Thakur T., and Meager M. R.. 2023. Neuropsychological Assessment. StatPearls: StatPearls Publishing. [PubMed] [Google Scholar]
  41. Sood, B. , Delaney‐Black V., Covington C., et al. 2001. “Prenatal Alcohol Exposure and Childhood Behavior at Age 6 to 7 Years: I. Dose–Response Effect.” Pediatrics 108: E34. [DOI] [PubMed] [Google Scholar]
  42. Walker, E. J. , Kirkham F. J., Stotesbury H., Dimitriou D., and Hood A. M.. 2023. “Tele‐Neuropsychological Assessment of Children and Young People: A Systematic Review.” Journal of Pediatric Neuropsychology 9: 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Wechsler, D. 2014. Wechsler Intelligence Scale for Children, Fifth Edition.
  44. Weintraub, S. , Dikmen S. S., Heaton R. K., et al. 2013. “Cognition Assessment Using the NIH Toolbox.” Neurology 80: S54–S64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Weyrauch, D. , Schwartz M., Hart B., Klug M. G., and Burd L.. 2017. “Comorbid Mental Disorders in Fetal Alcohol Spectrum Disorders: A Systematic Review.” Journal of Developmental and Behavioral Pediatrics 38: 283–291. [DOI] [PubMed] [Google Scholar]
  46. Wozniak, J. R. , Riley E. P., and Charness M. E.. 2019. “Clinical Presentation, Diagnosis, and Management of Fetal Alcohol Spectrum Disorder.” Lancet Neurology 18: 760–770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Zahn, I. 2010. “Working With Unbalanced Cell Sizes in Multiple Regression With Categorical Predictors.” Education 112: 1–18. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All or part of this work was done in conjunction with the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD, https://doi.org/10.5967/ntw9‐h991), which is funded by grants from the National Institute on Alcohol Abuse and Alcoholism (NIAAA). Additional information about CIFASD, including about data sharing opportunities, can be found at https://www.cifasd.org. Data from this study are part of Phase 4 (2017‐2022) and Phase 5 (2022‐2027).

Articles from Alcohol, Clinical & Experimental Research are provided here courtesy of Wiley

RESOURCES