Many individuals living with severe mental illness, such as schizophrenia, present cognitive deficits and reasoning biases negatively impacting clinical and functional trajectories. Remote cognitive assessment presents many opportunities for advancing research and treatment but has yet to be widely used in psychiatric populations. We conducted a scoping review of remote cognitive assessment in severe mental illness to provide an overview of available measures and guide best practices. Overall, 34 studies (n = 20,813 clinical participants) were reviewed and remote measures, psychometrics, facilitators, barriers, and future directions were synthesized using a logic model. We identified 82 measures assessing cognition in severe mental illness across 11 cognitive domains and four device platforms. Remote measures were generally comparable to traditional versions, though psychometric properties were infrequently reported. Facilitators included standardized procedures and wider recruitment, whereas barriers included imprecise measure adaptations, technology inaccessibility, low patient engagement, and poor digital literacy. Our review identified several remote cognitive measures in psychiatry across all cognitive domains. However, there is a need for more rigorous validation of these measures and consideration of potentially influential factors, such as sex and gender. We provide recommendations for conducting remote cognitive assessment in psychiatry and fostering high-quality research using digital technologies.
Cognitive impairment is a core feature of psychiatric illness, particularly schizophrenia and related disorders1,2. Robust cognitive deficits are observed in several cognitive domains in schizophrenia, including memory, attention, and executive function3,4,5. Less well-known cognitive symptoms in schizophrenia are cognitive biases, which are errors in judgment or interpretation that affect decision-making (e.g., jumping to conclusions, confirmation bias) and contribute to symptoms6,7,8. Both traditional cognitive impairments and elevated cognitive biases are rooted in neurobiology9,10 and affect many diagnosed with mental illness11,12,13, negatively impacting clinical and functional trajectories6,14. Cognitive assessments are essential in guiding treatment planning and, thus, proper measurement of both cognitive capacity and cognitive biases is fundamental to improve overall patient cognitive health and outcomes.
Remote cognitive assessments outside the clinic or laboratory have become a necessity in the context of the COVID-19 pandemic, which has hindered mental health initiatives in both research and clinical settings worldwide15,16. Yet, it also provides a rare opportunity for researchers and clinicians to draw from—and contribute to—the growing literature on remote digital technologies in psychiatry. Digital technology promoting mental health research and practice, or e-mental health, has become prevalent worldwide and can improve the implementation of evidence-based practice17,18. Most individuals with schizophrenia19 and first-episode psychosis20 have access to a computer, smartphone, or tablet and growing research supports the use, acceptability, feasibility, and efficacy of digital technologies in psychiatry21,22,23,24. Digital cognitive assessments are also being increasingly developed for these devices, with recent reviews suggesting they are feasible and reliable measures of cognition25,26,27.
Remote cognitive assessments provide many opportunities to advance research and treatment in severe mental illness, particularly schizophrenia-spectrum disorders. As they are typically digital measures, remote assessments can benefit from advances in the field of computerized neuropsychological assessment (e.g., ref. 28) as evidenced more broadly by the InterOrganizational Practice Committee guidelines for teleneuropsychology29. Remote assessments also offer the same advantages as computerized measures, including increased precision, standardized testing, and automated scoring25,30,31. Moreover, they enable the recruitment of larger and more diverse samples (e.g., from rural and remote areas) and of individuals who might have practical (e.g., cost, transportation) or symptomatic (e.g., social avoidance, paranoia) issues that make in-person attendance difficult. Assessments using tablets and smartphones have added benefits in that they can more easily be completed remotely at any time and in any geographic location25,32 and can provide data on additional dynamic variables (e.g., environment data, sleep quality, mood, level of exercise, etc.) for a broader assessment of cognition25.
There is an urgent need to verify that remote cognitive assessments provide valid assessments of cognitive capacity and cognitive biases in severe mental illness. Although recent reviews support the use of digital cognitive assessments in psychiatric populations, delivery in remote settings is not yet common25,26,27. Consequently, many researchers and clinicians are rapidly embarking on this path with little empirical evidence to provide guidance. The purpose of this scoping review is to provide an overview of the literature on remote cognitive assessment in severe mental illness. We focus on remote assessments in psychiatric illnesses rather than broad digital measures or remote measures in the general population given the great potential for remote assessments to drive research and treatment in this population25,26. We opted for a scoping review as they are designed to address broad, overarching research questions within a systematic review framework33,34. Our main population of interest included individuals with severe mental illness (e.g., schizophrenia-spectrum disorders), though we did not exclude research involving other groups. Our objectives were to map the current literature, identify potential barriers and facilitators, and highlight knowledge gaps in remote cognitive assessment in severe mental illness. This review aims to provide insight into the currently available options for clinicians and researchers and encourage high-quality research on remote cognitive assessment in psychiatry during and beyond the COVID-19 pandemic.
Selection of sources of evidence
Figure 1 displays the PRISMA flowchart, combining the retrieved articles across the three literature searches. In the initial search, 24,516 references were identified, including one in press manuscript through a co-author (SG). After the removal of 1760 duplicates, titles and abstracts of 22,756 articles were randomly divided and screened by five reviewers. Of these, 57 articles were flagged as potentially relevant and full texts were screened. Upon full-text review, 32 additional articles were excluded due to not meeting one or more of the selection criteria. One additional article was identified through a reference list search. An updated search after 6 months yielded an additional 859 articles, five of which met inclusion criteria, with one additional article found through reference list search. A final updated search 3 months later yielded an additional 1124 articles (note: search updates were limited by year and overlapped with previous searches), two of which met inclusion criteria. Thus, 34 articles were included in the scoping review, including a narrative review of digital technology for remote cognitive assessment in psychiatry26, a commentary on remote digital cognitive assessment in schizophrenia25, and a systematic review on digital assessment of verbal memory in first-episode psychosis27. These three nonexperimental articles are incorporated only into the facilitators, barriers, and future directions sections of the logic model and the remaining articles informed all sections of the model.
Inter- and Intra-rater reliability
Inter-rater reliability (IRR) was high at start, midpoint and end of article selection and increased over time: IRR 1 = 0.95 (SE = 0.02, 95% CI = [0.92–0.98], p < 0.001, two-sided); IRR 2 = 0.97 (SE = 0.01, 95% CI = [0.94–1.00], p < 0.001, two-sided); IRR 3 = 0.98 (SE = 0.01, 95% CI = [0.96–1.00], p < 0.001, two-sided). Supplementary Table 2 displays the distribution of disagreements (initial rating compared to consensus) per rater over the three IRRs. The number of articles (out of 100 at each timepoint) with conflicting ratings between two or more raters was low and decreased over time: IRR 1 = 10/100, IRR 2 = 6/100, IRR 3 = 4/100. The mean number of conflicts was also low and decreased (IRR 1 = 3.20, SD = 2.59; IRR 2 = 2.60, SD = 1.52; IRR 3 = 1.40, SD = 1.67).
Characteristics and results of sources of evidence
Table 1 lists the 31 experimental articles selected for review (excluding the three review articles of the total selected 34 articles), along with primary characteristics (psychiatric diagnosis, sample size, remote platform, supervision, battery/measure assessed, and relevant cognitive domain). Full study characteristics are displayed in Supplementary Data, including sociodemographics (sample size, control group, age ranges, sex ratios, country, language), measure characteristics (study setting, researcher presence and title, license type, measure type, duration, alternate forms), psychometric properties, and sex-related findings. Selected articles were published between 2009 and 2021, though most (82.35%) were published within the past 5 years.
Synthesis of results: logic model
The final logic model is presented in Fig. 2. The central panel includes 82 remote cognitive measures divided into 11 cognitive domains. The most assessed domains were speed of processing, working memory, reasoning, and executive function, whereas subjective cognition included only a single reviewed measure. For each measure, we illustrate which platform(s) were used (videoconference, web browser, tablet, and smartphone, in normal, bold, underline, and italic font, respectively) and whether the assessment was tested in a laboratory setting (white circle), remotely (black circle) or both (white and black circle). Briefly, two studies tested their measures using videoconferencing, 16 via web browser, two with a tablet, and nine with smartphones. Only one study35 reported their remote assessment could be performed on two platforms (i.e., tablet and web browser) through several used web-based measures that could likely be used on several platforms (e.g., web, smartphone, tablet). In total, six studies included remote measures that were completed in a laboratory setting, 23 were done remotely, and two used both settings.
The upper circles of the logic model summarize reported reliability, sensitivity/specificity, construct validity, and criterion validity of the reviewed measures, detailed in Supplementary Data. For each cognitive domain, we report the number of times a given psychometric was evaluated over the total number of times it was measured across studies. Next to each total, we summarize the reported psychometric properties as either low (L), moderate (M), or high (H) and invite the reader to consult Supplementary Data for detailed findings. Reliability includes estimates of internal consistency, test-retest evaluations, and intraclass correlations. Sensitivity and specificity respectively refer to the ability of the reviewed measure to identify those with and without impairments. Construct validity includes correlations with comparison measures (e.g., pen-and-paper versions) and correlations between human and automated scoring. Criterion validity includes correlations between the reviewed measures and outcomes, such as sociodemographics, symptoms, and functioning. Construct validity was most frequently assessed irrespective of the cognitive domain, whereas reliability was assessed least frequently. Overall, we observe that, for measures in which psychometric properties are assessed, remote measures were generally as reliable, sensitive, and valid as traditional measures. One exception was social cognition, which showed poor discriminatory power in one study36 and low to moderate correlations with traditional measures (see Supplementary Data).
The lower panels of the logic model outline thematically defined barriers and facilitators to the development and implementation of remote cognitive assessment as well as proposed improvements and avenues for future research. For development, facilitators included incorporating standardized procedures, alternate measure versions, and using technology to mitigate potential barriers (e.g., preloading stimuli to limit internet connectivity issues). On the other hand, developmental barriers included confidentiality concerns, technology/system variability, imprecise measure adaptations, and the current lack of remote norms. For implementation, testing in a neutral setting, improving feasibility (reminders, user-friendly technology), and wider access to individuals living in rural regions have been identified as facilitators. Inversely, low participant engagement, symptom severity, limited digital literacy, poor technology accessibility, and potential access to outside help (e.g., through family members or the internet) have been identified as barriers. As for proposed improvements and future directions, the authors of reviewed studies highlighted the need for further psychometric validation, development of remote norms, and strategies to ensure digital security. There were also proposed improvements pertaining to the promotion of open-source options, optimization of collected data (detailed cognitive performance data and additional contextual variables, such as sleep and physical activity), and verification of diagnostic and cultural generalizability.
Sex and gender considerations
Given the well-documented sex differences in cognition and their relevance to psychiatric illness37,38, we sought to examine the role of sex and gender on remote assessment of cognitive capacity and cognitive biases. Approximately one-quarter of experimental studies (n = 9) reported on differences based on sex assigned at birth (male, female) and none on gender identity (e.g., non-binary, trans-, cis-, genderfluid). Sex and gender were often used interchangeably presumably in reference to sex assigned at birth. One study reported matching participants based on sex and used sex-corrected pen-and-paper norms39, one did not report explicit sex ratios40, and one included females only41. Those that reported on sex differences found that females displayed higher cognitive biases42 and lower performance on working memory43. Two articles described nonspecific sex differences36,44, and three found no sex-related performance40,45 or attrition46 differences (see Supplementary Data).
The present study provides a scoping review of the literature on remote assessment of cognitive capacity and cognitive biases in severe mental illness to map current knowledge and inform clinicians and researchers on best practices. In total, more than 26,000 articles were retrieved and 34 met our inclusion criteria. Identified measures generally showed acceptable psychometric properties, though these were assessed in less than half of reviewed studies. Facilitators and barriers to the development and implementation of remote cognitive assessment measures, as well as future research directions proposed by identified studies, provide clear considerations for future research and practice. This work brings together the current library of remote cognitive measures in psychiatry that researchers and clinicians may consult based on their needs, including cognitive domain, remote platform, and level of supervision required. Below we provide general recommendations and considerations to foster remote cognitive assessment in psychiatry.
Our scoping review did not identify a “gold-standard” remote battery for a comprehensive assessment of cognition in psychiatric populations. Moreover, there is currently no single cognitive battery, remote or otherwise, assessing both cognitive capacity and cognitive biases to provide an overall measure of cognitive health in severe mental illness. For cognitive capacity, the two most frequently used computerized cognitive batteries in psychiatric populations (CANTAB and CogState)47 did not emerge strongly in our review, suggesting they have not yet been adopted extensively in remote settings despite their potential for remote administration. Only one study41 used the CogState Brief Battery in a remote setting in a very large sample of nurses with elevated PTSD symptoms, though the generalizability of the results to other psychiatric samples remains in question. CANTAB was only used in a single study as an in-lab comparison measure35. Moreover, social cognition measures were restricted to emotion recognition tasks and tests of other domains of social cognition (e.g., theory of mind) are currently lacking. Notable comprehensive remote batteries that reported acceptable psychometric properties included the Brief Assessment of Cognition39, My Cognition Quotient35, Online Neurocognitive Assessments36, and Screen for Cognitive Assessment in Psychiatry48. Some individual tasks also showed valid, sensitive, and/or reliable remote administration, particularly the Jewel Trail Making Task from the mindLAMP smartphone application, used in three studies49,50,51.
Cognitive biases were primarily assessed using scales rather than tasks, which are more amenable to remote administration via online survey platforms. Importantly, most cognitive bias scales and all cognitive bias tasks identified were designed to address individual biases, such as jumping to conclusions45,52,53,54. The most general measure of cognitive biases identified was the Davos Assessment of Cognitive Biases Scale55, though it does not measure all biases reported in psychiatric disorders. Surprisingly, the well-known Cognitive Biases Questionnaire for Psychosis56 did not emerge in our review, suggesting it has yet to be used in remote settings with severe mental illness. Given the importance of cognitive biases in understanding and treating the symptoms of severe mental illness7, the development of a validated remote cognitive bias battery to complement the numerous batteries that exist to assess cognitive capacity is recommended.
Fundamentally, the question of which measure(s) to use depends on the cognitive domain(s) of interest and other pragmatic considerations (platform, duration, cost, etc). Comprehensive batteries would likely be most convenient for clinicians and for researchers interested in general measures of cognition across various domains. However, most of the available comprehensive cognitive batteries are proprietary (Supplementary Data) and thus incur significant costs and less flexibility for the user. Several open-source measures were available through online platforms, such as Inquisit Web or researcher-developed applications. There exist other promising experiment-sharing platforms (e.g., Pavlovia, Expyriment, CognitionLab), though, to our knowledge, these have yet to be tested remotely with psychiatric samples. Generally, these platforms require “picking and choosing” and/or developing cognitive measures and thus necessitate greater reflection on the objectives and cognitive measures of interest. True open-source alternatives, in which the task’s source code is fully accessible are also available for some measures, or reportedly available from the authors. These initiatives would likely be of greater interest to cognitive scientists.
While this review illustrates that remote cognitive assessment is feasible with psychiatric populations, most studies strongly recommended further validation of existing remote measures, development of additional measures, and remote norms. Remote norms were not reported in the identified studies, despite the potential for remote studies to collect data from large and diverse samples and the growing number of computerized batteries with normative data (e.g., refs. 57,58,59,60). Only one selected study assessed whether in-lab computerized scores were comparable to pen-and-paper norms, finding that modifications were necessary for some subtests of the Brief Assessment of Cognition39. Thus, normative data derived from in-person assessments might not be applicable to computerized or remote versions of all cognitive tests. The development of remote norms would greatly facilitate remote cognitive assessment and allow for improved comparisons between studies. However, this poses several challenges. Notably, comparable in-person normative data are not available for all tests, particularly for measures of cognitive biases. In addition, the nature of remote assessment occurring outside the laboratory naturally reduces researchers’ control over environmental confounds that could affect test performance. Future development of remote normative data and guidelines for such norms should address these potential issues.
Additional quality considerations should be made during both the development and implementation of a new cognitive task or study. In terms of development, identified studies strongly encouraged using standardized and automated procedures, including instructions and scoring, to reduce variability and human error. Moreover, eliminating the need for a synchronous internet connection (e.g., preloading cognitive stimuli and allowing test results to be uploaded asynchronously) can mitigate potential issues with internet connectivity. Adaptation of certain pen-and-paper measures to remote computerized software also presents a major challenge to validity and feasibility, particularly for those measures that involve writing or motor skills, and pen-and-paper norms may be inaccurate in these cases. The choice of remote platform (web, tablet, smartphone, videoconference) or multi-platform options should also be carefully evaluated, as platforms vary in terms of functionality (e.g., touch screen ability) and other parameters (e.g., screen size, computational power) that can affect performance. It is also imperative to ensure that collected data corresponds to high ethical standards in terms of security and privacy, including transparency, confidentiality, data safeguarding, and avoiding superfluous data collection61,62. Finally, when implementing cognitive assessments in remote settings, participants’ digital competence, symptom severity, and potential environmental distractors should be considered, all of which can affect performance over and above cognitive impairments. Reminder notifications, standardized instructions, practice, and remote monitoring may limit these potential issues.
Future remote studies should prioritize larger samples, standardization of instructions and environment, where possible, broader data collection (e.g., environmental data, sleep quality, mood, level of exercise, etc.) and wider recruitment (e.g., remote and rural areas) to allow for the development of norms and to assess potential sociodemographic factors (sex, gender, race, education, etc.) and diagnostic and cultural generalizability. Development and validation of additional remote measures of both cognitive capacity and cognitive biases would also bring us closer to developing an overall battery of cognitive health for those with psychiatric disorders.
Quality remote cognitive assessments have strong implications for remote cognitive interventions in psychiatry. Effective cognitive interventions are available for both cognitive capacity (e.g., cognitive remediation therapy)63,64,65,66 and cognitive biases (e.g., metacognitive training, cognitive behavioral therapy for psychosis)6,67,68. In a complimentary review and meta-analysis on the efficacy of virtual evidence-based psychosocial interventions for schizophrenia-spectrum disorders69, 11 studies met inclusion criteria for virtually-delivered cognitive remediation. Six of these were included in a meta-analysis showing moderate effects on neurocognition (Hedges g = 0.35) and functioning (g = 0.33), similar to in-person interventions66. These initial results on efficacy are promising for virtual adaptations of existing interventions and encourage the development of new programs specifically designed for virtual delivery. For example, patient-tailored remote interventions following a preliminary remote cognitive assessment would integrate personalized treatment and broad accessibility.
The current study presents several strengths. First, it is a broad scoping review of remote measures of both cognitive capacity and cognitive biases in severe mental illness designed to address an urgent need given the COVID-19 pandemic. Second, it involves rigorous methodological procedures including randomization, repeated inter-rater reliability, extensive quality control, and iterative data synthesis. Third, the search was updated after six and nine months given the rapidly evolving literature in this domain. Finally, data extraction was comprehensive and included several characteristics (e.g., diagnosis, setting, researcher presence, platform, duration, alternate forms, licensing, cognitive domain, psychometric properties) that will assist researchers and clinicians in their choice of remote measures.
A potential limitation of this study is that the search strategy, which was focused on severe mental illness, may not have captured all articles assessing remote cognition in other psychiatric disorders, though several were identified, and reference lists were also checked. Additionally, we did not calculate quality scores for included studies. Contrary to systematic literature reviews, a critical appraisal of sources of evidence is not generally indicated for scoping reviews, which are meant to be broadly inclusive of the literature70. Third, despite our best efforts, our review may have missed some findings from unpublished studies and ongoing investigations. This is particularly relevant given the present surge in remote research due to the COVID-19 pandemic and is illustrated by the eight additional sources of evidence identified in our updated searches. There are also many additional remote cognitive measures and batteries that were identified during the review process, but these had not yet been tested in populations with severe mental illness and were outside the scope of this review. Lastly, our domain classifications may not accurately represent all cognitive function(s) assessed by a given measure. However, this classification was developed using an iterative process until consensus was reached by the three lead authors and was reviewed and approved by the remaining authors, all of whom are experienced in the field.
At present, researchers and clinicians in psychiatry can choose from a vast selection of remote cognitive measures assessing many cognitive domains through various remote platforms. However, there is an urgent need for more rigorous validation of these measures and for a stronger consideration of influential factors, such as sex and gender differences and cultural diversity. Remote cognitive assessment is necessary given the current climate but also has the potential to propel the field of cognitive psychiatry forward. In conclusion, this review provides clinicians and researchers with a comprehensive list of remote cognitive assessment measures as well as insight into methodological and practical considerations that may serve as a first step in the development of guidelines for remote cognitive assessment in severe mental illness.
Protocol and registration
The review protocol was preregistered on the Open Science Framework: https://osf.io/cbzq8 (Registration 10.17605/OSF.IO/CBZQ8) and followed the PRISMA extension for scoping reviews71 (see Supplementary Table 3 for PRISMA checklist) and the Joanna Briggs Institute guidance on conducting systematic scoping reviews34,70,72.
Search strategy and selection criteria
A comprehensive literature search was conducted on May 11, 2020 and updated on November 11, 2020, and February 4, 2021 using OVID (MEDLINE, PsycInfo, and EMBASE) and EBSCO (CINAHL) databases. The following keywords were used: (schizophreni* OR psychosis OR psychoses OR psychotic* OR severe mental illness) AND (cogniti* OR neuropsych* OR bias* OR reason*) AND (remote* OR online* OR mobile* OR digital*) AND (assessment OR evaluat* OR test* OR measure*). The search was limited to articles in either English or French from any publication year. Evidence sources included peer-reviewed research articles, reviews, and letters to the editor, excluding books and conference abstracts. Repositories of tests and measures were searched (PsycTESTS, Health and Psychosocial Instruments, Mental Measurements Yearbook), experts were contacted for unpublished findings, and reference lists of selected articles were examined for additional studies.
Retrieved articles were combined in Endnote software, and in a first pass, duplicates were excluded automatically by comparing Author, Year, Title, and Journal fields. Duplicates based on all possible other combinations of these fields were produced and checked manually. The remaining articles were randomized for initial screening based on title and abstract. Due to the urgent nature of this review, five raters were assigned to screen the remaining de-duplicated articles, with each rater screening approximately one-fifth of the total number of de-duplicated articles. Raters assigned each article one of three possible ratings: include, questionable, exclude. To determine whether questionable articles should be included or excluded, full texts were reviewed according to the study inclusion criteria and a consensus was reached by the research team.
Article screening was based on the following eligibility criteria: (a) peer-reviewed; (b) included individuals with a diagnosis involving severe mental illness (e.g., schizophrenia-spectrum disorders); and (c) reported on the remote assessment of cognitive capacity and/or cognitive biases. During article selection, we recognized that several articles included a broad range of diagnostic groups (e.g., anxiety, depression, OCD) and we included these conditions to maintain a broader scope. In addition, many articles assessed remote cognitive tasks in a laboratory setting (e.g., comparison with a standard pen-and-paper battery). In order to include these articles, which were not technically remote, while not including all articles reporting on computerized cognitive assessment in psychiatry, we included these on a case-by-case basis, and the inclusion of articles determined via consensus. Selected articles were then retrieved for full-text screening and data extraction of included articles.
Given that articles were screened by different raters, rather than by all raters, we assessed inter-rater reliability (IRR) by having all raters assign ratings to three samples of 100 articles at the start, midpoint, and end of article selection, as in previous research73. IRR was calculated using Gwet’s AC1 statistic74 via the R AgreeStat package to account for the kappa paradox, in which unusually low agreement statistics are produced when there is a skewed distribution of ratings (e.g., many excluded articles)75,76,77 (see Supplementary Table 1 for a demonstration with the current data). Following each IRR timepoint, raters produced a consensus for any inconsistent ratings. Intra-rater reliability was also assessed across IRR timepoints73 by comparing each rater’s accuracy relative to consensus.
Data extraction was performed on selected articles according to a pre-developed form, which was tested and fine-tuned with one exemplar article by the lead author. Articles were randomized for data extraction across three independent raters. Data extraction was quality controlled by authors K.M.L., G.S., and D.R.-C. by randomly selecting six articles (10% of articles originally extracted) and re-extracting the data. Data extraction included the following predetermined variables: bibliographic data (authors, year, title, abstract), study characteristics (aims, design, country, setting, researcher presence/title, sample size, psychiatric diagnosis, mean age, age range, sex/gender ratio), description of remote assessment methods (remote/comparison measure(s), battery, remote platform, developer, language, duration, alternate forms, availability of norms), main findings, sex/gender findings, psychometric properties (reliability, sensitivity/specificity, construct validity, criterion validity), facilitators, barriers, and future directions.
Synthesis of results
Data were synthesized and illustrated using the logic model methodology, following the W. K. Kellogg Foundation guidelines78 and previous research79,80. This flexible method uses an iterative approach to identify and illustrate thematic categories and the putative underlying links to portray complex relationships81,82. In this study, the logic model was used to classify cognitive measures into domains (speed of processing, attention and vigilance, working memory, verbal learning and memory, visual learning and memory, reasoning and executive function, social cognition, verbal fluency, cognitive bias, subjective cognition, and IQ), expanded from the MATRICS83 classification. The logic model also outlines psychometric properties, facilitators, barriers, and future directions identified.
Logic model development
Cognitive measures were categorized into cognitive domains, which were inspired by the MATRICS83 classification: speed of processing, attention and vigilance, working memory, verbal learning and memory, visual learning and memory, reasoning and problem solving, and social cognition. We added verbal fluency, cognitive bias, subjective cognition, and IQ domains, to account for identified measures which did not fit within the MATRICS domains. We initially selected the MATRICS classification as it provides a well-known framework for cognitive impairment in schizophrenia, which was our primary population of interest and the group assessed in most studies. In addition, the MATRICS domain of reasoning and problem solving was relabeled as “reasoning and executive function” in order to include additional measures of executive functioning (e.g., inhibitory control) without creating a separate domain. Notably, several measures tap into additional domains reported in the literature (e.g., visuomotor processing) or recruit additional cognitive processes that fall into other identified domains (e.g., speed of processing measures also require attention). In the current review, a given measure’s primary cognitive domain is reported and was determined through consensus.
Differences between draft and final logic model
Prior to data extraction, we developed a draft logic model (Supplementary Fig. 1). The final logic model (Fig. 2) was developed through an iterative process by the three lead authors and was reviewed and approved by the remaining authors. Differences between the draft and final logic models are outlined below. In both models, the identified remote cognitive measures, relevant cognitive domains, and procedural characteristics are displayed in the middle panel. Psychometric properties are located in the upper circles and facilitators, barriers, and future directions in the lower sections.
The draft logic model categorized the identified remote cognitive assessment measures categorized by MATRICS cognitive domain (speed of processing, attention/vigilance, working memory, verbal memory and learning, visual memory and learning, reasoning and problem solving, social cognition)83 with the addition of a cognitive bias domain. The draft model also outlined utilized procedures (setting, platform, researcher presence/title, duration, material, cost), psychometric properties (reliability, sensitivity/specificity, concurrent validity, predictive validity), facilitators, barriers, improvements/future research, and recommendations. During data extraction, we decided to report on the license type (proprietary, open-source) of a given measure/battery, rather than cost, as this was not readily available. To simplify the presentation, the final logic model reports only the platform and location of testing with the other procedures relegated to Supplementary Data. We also renamed concurrent and predictive validity to construct and criterion validity, respectively, to emphasize the wider breadth of psychometric properties that were available. Facilitators, barriers, improvements/future research remained unchanged from the draft to the final logic model.
Inter-rater reliability was calculated with the AgreeStat package in R, available at: https://agreestat.com/software/r/new/agree.coeff3.raw.r.
Kahn, R. S. & Keefe, R. S. Schizophrenia is a cognitive illness: time for a change in focus. JAMA Psychiatry 70, 1107–1112 (2013).
Solé, B. et al. Cognitive impairment in bipolar disorder: treatment and prevention strategies. Int. J. Neuropsychopharmacol. 20, 670–680 (2017).
Bora, E., Binnur Akdede, B. & Alptekin, K. Neurocognitive impairment in deficit and non-deficit schizophrenia: a meta-analysis. Psychol. Med. 47, 2401–2413 (2017).
Rock, P., Roiser, J., Riedel, W. & Blackwell, A. Cognitive impairment in depression: a systematic review and meta-analysis. Psychol. Med. 44, 2029 (2014).
Vöhringer, P. A. et al. Cognitive impairment in bipolar disorder and schizophrenia: a systematic review. Front. Psychiatry 4, 87 (2013).
Sauvé, G., Lavigne, K. M., Pochiet, G., Brodeur, M. & Lepage, M. Efficacy of psychological interventions targeting cognitive biases in schizophrenia: a systematic review and meta-analysis. Clin. Psychol. Rev. 78, 101854 (2020).
Broyd, A., Balzan, R. P., Woodward, T. S. & Allen, P. Dopamine, cognitive biases and assessment of certainty: a neurocognitive model of delusions. Clin. Psychol. Rev. 54, 96–106 (2017).
Everaert, J., Koster, E. H. W. & Derakshan, N. The combined cognitive bias hypothesis in depression. Clin. Psychol. Rev. 32, 413–424 (2012).
Kelly, S. et al. Neural correlates of cognitive deficits across developmental phases of schizophrenia. Neurobiol. Dis. 131, 104353 (2018).
Lavigne, K. M., Menon, M. & Woodward, T. S. Functional brain networks underlying evidence integration and delusions in schizophrenia. Schizophr. Bull. 46, 175–183 (2020).
Bora, E. & Pantelis, C. Meta-analysis of cognitive impairment in first-episode bipolar disorder: comparison with first-episode schizophrenia and healthy controls. Schizophr. Bull. 41, 1095–1104 (2015).
McCleery, A. & Nuechterlein, K. H. Cognitive impairment in psychotic illness: prevalence, profile of impairment, developmental course, and treatment considerations. Dialogues Clin. Neurosci. 21, 239–248 (2019).
Sauve, G., Malla, A., Joober, R., Brodeur, M. B. & Lepage, M. Comparing cognitive clusters across first- and multiple-episode of psychosis. Psychiatry Res. 269, 707–718 (2018).
Lepage, M., Bodnar, M. & Bowie, C. R. Neurocognition: clinical and functional outcomes in schizophrenia. Can. J. Psychiatry 59, 5–12 (2014).
Türközer, H. B. & Öngür, D. A projection for psychiatry in the post-COVID-19 era: potential trends, challenges, and directions. Mol. Psychiatry 25, 2214–2219 (2020).
Öngür, D., Perlis, R. & Goff, D. Psychiatry and COVID-19. JAMA 324, 1149–1150 (2020).
Lal, S. E-mental health: promising advancements in policy, research, and practice. Healthc. Manage. Forum 32, 56–62 (2019).
Wise, J. et al. WPA position statement on e-mental health. World Psychiatric Association (2017).
Gay, K., Torous, J., Joseph, A., Pandya, A. & Duckworth, K. Digital technology use among individuals with schizophrenia: results of an online survey. JMIR Ment. Health 3, e15 (2016).
Abdel-Baki, A., Lal, S., D.-Charron, O., Stip, E. & Kara, N. Understanding access and use of technology among youth with first-episode psychosis to inform the development of technology-enabled therapeutic interventions. Early Interv. Psychiatry 11, 72–76 (2017).
Sin, J. et al. Digital interventions for screening and treating common mental disorders or symptoms of common mental illness in adults: systematic review and meta-analysis. J. Med. Internet Res. 22, e20581 (2020).
Lal, S. et al. Preferences of young adults with first-episode psychosis for receiving specialized mental health services using technology: a survey study. JMIR Ment. Health 2, e18 (2015).
Merchant, R., Torous, J., Rodriguez-Villa, E. & Naslund, J. A. Digital technology for management of severe mental disorders in low-income and middle-income countries. Curr. Opin. Psychiatry 33, 501–507 (2020).
Lecomte, T. et al. Mobile apps for mental health issues: meta-review of meta-analyses. JMIR Mhealth Uhealth 8, e17458 (2020).
Guimond, S., Keshavan, M. S. & Torous, J. B. Towards remote digital phenotyping of cognition in schizophrenia. Schizophr. Res. 208, 36–38 (2019).
Hays, R. et al. Assessing cognition outside of the clinic: smartphones and sensors for cognitive assessment across diverse psychiatric disorders. Psychiatr. Clin. North Am. 42, 611–625 (2019).
Kilciksiz, C. M., Keefe, R., Benoit, J., Ongur, D. & Torous, J. Verbal memory measurement towards digital perspectives in first-episode psychosis: a review. Schizophr. Res. Cogn. 21, 100177 (2020).
Bauer, R. M. et al. Computerized neuropsychological assessment devices: joint position paper of the American Academy of Clinical Neuropsychology and the National Academy of Neuropsychology. Arch. Clin. Neuropsychol. 27, 362–373 (2012).
Bilder, R. M. et al. InterOrganizational practice committee recommendations/guidance for teleneuropsychology (TeleNP) in response to the COVID-19 pandemic. Arch. Clin. Neuropsychol. 34, 1314–1334 (2020).
Schatz, P. & Browndyke, J. Applications of computer-based neuropsychological assessment. J. Head Trauma Rehabil. 17, 395–410 (2002).
Langenecker, S. A. et al. The sensitivity and psychometric properties of a brief computer-based cognitive screening battery in a depression clinic. Psychiatry Res. 152, 143–154 (2007).
Koo, B. M. & Vizer, L. M. Mobile technology for cognitive assessment of older adults: a scoping review. Innov. Aging 3, igy038 (2019).
Munn, Z. et al. Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Med. Res. Methodol. 18, 143 (2018).
Khalil, H. et al. Conducting high quality scoping reviews-challenges and solutions. J. Clin. Epidemiol 130, 156–160 (2021).
Domen, A. C., van de Weijer, S. C. F., Jaspers, M. W., Denys, D. & Nieman, D. H. The validation of a new online cognitive assessment tool: the MyCognition Quotient. Int. J. Methods Psychiatric Res. 28, e1775 (2019).
Biagianti, B. et al. Development and testing of a web-based battery to remotely assess cognitive health in individuals with schizophrenia. Schizophr. Res. 208, 250–257 (2019).
Buck, G. et al. Sex differences in verbal memory predict functioning through negative symptoms in early psychosis. Schizophr. Bull 46, 1587–1595 (2020).
Vaskinn, A. et al. Sex differences in neuropsychological performance and social functioning in schizophrenia and bipolar disorder. Neuropsychology 25, 499–510 (2011).
Atkins, A. S. et al. Validation of the tablet-administered brief assessment of cognition (BAC App). Schizophr. Res. 181, 100–106 (2017).
Pop-Jordanova, N., Loleska, S. & Loleski, M. Originally adapted mobile application used for neuropsychiatric patients. Prilozi 39, 75–81 (2018).
Sumner, J. A. et al. Posttraumatic stress disorder symptoms and cognitive function in a large cohort of middle-aged women. Depress. Anxiety 34, 356–366 (2017).
Metel, D. et al. Mediating role of cognitive biases, resilience and depressive symptoms in the relationship between childhood trauma and psychotic-like experiences in young adults. Early Interven. Psychiatry 14, 87–96 (2020).
Eraydin, I. E. et al. Investigating the relationship between age of onset of depressive disorder and cognitive function. Int. J. Geriatric Psychiatry 34, 38–46 (2019).
Miegel, F., Jelinek, L. & Moritz, S. Dysfunctional beliefs in patients with obsessive-compulsive disorder and depression as assessed with the Beliefs Questionnaire (BQ). Psychiatry Res. 272, 265–274 (2019).
Moritz, S. et al. Prolonged rather than hasty decision-making in schizophrenia using the box task. Must we rethink the jumping to conclusions account of paranoia?. Schizophr. Res 222, 222–202 (2020).
Hung, S. et al. Smartphone-based ecological momentary assessment for Chinese patients with depression: an exploratory study in Taiwan. Asian J. Psychiatry 23, 131–136 (2016).
Bakkour, N. et al. Systematic review of appropriate cognitive assessment instruments used in clinical trials of schizophrenia, major depressive disorder and bipolar disorder. Psychiatry Res. 216, 291–302 (2014).
Bernardo-Ramos, M., Franco-Martin, M. A. & Soto-Perez, F. Cyber-Neuropsychology: application of new technologies in neuropsychological evaluation. Actas Esp. Psiquiatria 40, 308–314 (2012).
Schvetz, C., Gu, F., Drodge, J., Torous, J. & Guimond, S. Validation of an ecological momentary assessment to measure processing speed and executive function in schizophrenia. NPJ Schizophr 7, 1–9 (2021).
Liu, G., Henson, P., Keshavan, M., Pekka-Onnela, J. & Torous, J. Assessing the potential of longitudinal smartphone based cognitive assessment in schizophrenia: a naturalistic pilot study. Schizophr. Res. 17, 100144 (2019).
Hays, R., Keshavan, M., Wisniewski, H. & Torous, J. Deriving symptom networks from digital phenotyping data in serious mental illness. BJPsych Open 6, e135 (2020).
Ludtke, T., Kriston, L., Schroder, J., Lincoln, T. M. & Moritz, S. Negative affect and a fluctuating jumping to conclusions bias predict subsequent paranoia in daily life: an online experience sampling study. J. Behav. Ther. Exp. Psychiatry 56, 106–112 (2017).
Moritz, S. et al. The benefits of doubt: cognitive bias correction reduces hasty decision-making in schizophrenia. Cognit. Ther. Res. 39, 627–635 (2015).
Moritz, S. et al. Metacognition-augmented cognitive remediation training reduces jumping to conclusions and overconfidence but not neurocognitive deficits in psychosis. Front. Psychol. 6, 1048 (2015).
van der Gaag, M. et al. Development of the Davos assessment of cognitive biases scale (DACOBS). Schizophr. Res. 144, 63–71 (2013).
Peters, E. R. et al. Cognitive biases questionnaire for psychosis. Schizophr. Bull. 40, 300–313 (2014).
Germine, L. et al. Is the Web as good as the lab? Comparable performance from Web and lab in cognitive/perceptual experiments. Psychon. Bull. Rev. 19, 847–857 (2012).
Cromer, J. A. et al. Comparison of cognitive performance on the Cogstate brief battery when taken in-clinic, in-group, and unsupervised. Clin. Neuropsychol. 29, 542–558 (2015).
Backx, R., Skirrow, C., Dente, P., Barnett, J. H. & Cormack, F. K. Comparing web-based and lab-based cognitive assessment using the Cambridge Neuropsychological Test Automated Battery: a within-subjects counterbalanced study. J. Med. Internet Res. 22, e16792 (2020).
Silverstein, S. M. et al. Development and validation of a World-Wide-Web-based neurocognitive assessment battery: WebNeuro. Behav. Res. Methods 39, 940–949 (2007).
Gelinas, L., Morrell, W., White, S. A. & Bierer, B. E. Navigating the ethics of remote research data collection. Clin. Trials 18, 606–614 (2021).
Hensen, B. et al. Remote data collection for public health research in a COVID-19 era: ethical implications, challenges and opportunities. Health Policy Planning 36, 360–368 (2021).
Bowie, C. R., Grossman, M., Gupta, M., Holshausen, K. & Best, M. W. Action-based cognitive remediation for individuals with serious mental illnesses: effects of real-world simulations and goal setting on functional and vocational outcomes. Psychiatr. Rehabil. J. 40, 53–60 (2017).
Bowie, C. R., McGurk, S. R., Mausbach, B., Patterson, T. L. & Harvey, P. D. Combined cognitive remediation and functional skills training for schizophrenia: effects on cognition, functional competence, and real-world behavior. Am. J. Psychiatry 169, 710–718 (2012).
Cella, M., Preti, A., Edwards, C., Dow, T. & Wykes, T. Cognitive remediation for negative symptoms of schizophrenia: a network meta-analysis. Clin. Psychol. Rev. 52, 43–51 (2017).
Wykes, T., Huddy, V., Cellard, C., McGurk, S. R. & Czobor, P. A meta-analysis of cognitive remediation for schizophrenia: methodology and effect sizes. Am. J. Psychiatry 168, 472–485 (2011).
Eichner, C. & Berna, F. Acceptance and efficacy of metacognitive training (MCT) on positive symptoms and delusions in patients with schizophrenia: a meta-analysis taking into account important moderators. Schizophr. Bull. 42, 952–962 (2016).
Moritz, S., Klein, J. P., Lysaker, P. H. & Mehl, S. Metacognitive and cognitive-behavioral interventions for psychosis: new developments. Dialogues Clin. Neurosci. 21, 309–317 (2019).
Jagtap, S., Romanowska, S., Leibovitz, T., Onno, K. A., Burhan, A. M., & Best, M. W. Can cognitive remediation therapy be delivered remotely? A review examining feasibility and acceptability of remote interventions. Schizophrenia Research: Cognition, 100238. https://doi.org/10.1016/j.scog.2022.100238 (2022).
Peters, M. D. J. et al. Updated methodological guidance for the conduct of scoping reviews. JBI Evid. Synth 18, 2119–2126 (2020).
Tricco, A. C. et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann. Intern. Med. 169, 467–473 (2018).
Peters, M. D. J. et al. Chapter 11: Scoping Reviews. In JBI Manual for Evidence Synthesis (eds Aromataris, E & Munn, Z.) JBI. (2020).
Belur, J., Tompson, L., Thornton, A. & Simon, M. Interrater reliability in systematic review methodology: exploring variation in coder decision-making. Sociol. Methods Res 50, 837–865 (2021).
Gwet, K. L. Computing inter-rater reliability and its variance in the presence of high agreement. Br. J. Math. Stat. Psychol. 61, 29–48 (2008).
Feinstein, A. R. & Cicchetti, D. V. High agreement but low kappa: I. The problems of two paradoxes. J. Clin. Epidemiol. 43, 543–549 (1990).
Wongpakaran, N., Wongpakaran, T., Wedding, D. & Gwet, K. L. A comparison of Cohen’s Kappa and Gwet’s AC1 when calculating inter-rater reliability coefficients: a study conducted with personality disorder samples. BMC Med. Res. Methodol. 13, 61 (2013).
Zec, S., Soriani, N., Comoretto, R. & Baldi, I. High agreement and high prevalence: the paradox of Cohen’s Kappa. Open Nurs. J. 11, 211–218 (2017).
W.K. Kellogg Foundation Team. Logic Model Development Guide https://www.wkkf.org/resource-directory/resources/2004/01/logic-model-development-guide W.K. Kellogg Foundation. (2004).
Baxter, S. K. et al. Using logic model methods in systematic review synthesis: describing complex pathways in referral management interventions. BMC Med. Res. Methodol. 14, 62 (2014).
Winsper, C., Crawford-Docherty, A., Weich, S., Fenton, S. J. & Singh, S. P. How do recovery-oriented interventions contribute to personal mental health recovery? A systematic review and logic model. Clin. Psychol. Rev. 76, 101815 (2020).
Chen, H.-T. Theory-driven Evaluations (Sage Publications, Inc, 1990).
Conrad, K. J., Randolph, F. L., Kirby, M. W. J. & Bebout, R. R. Creating and using logic models. Alcohol. Treat. Q. 17, 17–31 (1999).
Nuechterlein, K. H. et al. The MATRICS Consensus Cognitive Battery, part 1: test selection, reliability, and validity. Am. J. Psychiatry 165, 203–213 (2008).
Lavigne, K. M., Sauve, G. & Lepage, M. Remote cognitive assessment in severe mental illness: a scoping review. https://osf.io/wh6vt/ (2020).
Biagianti, B., Schlosser, D., Nahum, M., Woolley, J. & Vinogradov, S. Creating live interactions to mitigate barriers (CLIMB): a mobile intervention to improve social functioning in people with chronic psychotic disorders. JMIR Mental Health 3, e52 (2016).
Depp, C. A. et al. Ecological momentary facial emotion recognition in psychotic disorders. Psychol. Med. 1–9, (2021).
Dupuy, M. et al. Mobile cognitive testing in patients with schizophrenia: a controlled study of feasibility and validity. Journal de Therapie Comportementale et Cognitive 28, 204–213 (2018).
Holmlund, T. B. et al. Applying speech technologies to assess verbal memory in patients with serious mental illness. npj Digit. Med. 3, 33 (2020).
Kuhn, S., Berna, F., Ludtke, T., Gallinat, J. & Moritz, S. Fighting depression: action video game play may reduce rumination and increase subjective and objective cognition in depressed patients. Front. Psychol. 9, 129 (2018).
Moritz, S. & Jelinek, L. Inversion of the “unrealistic optimism” bias contributes to overestimation of threat in obsessive-compulsive disorder. Behav. Cogn. Psychother. 37, 179–193 (2009).
Moritz, S. et al. Repetition is good? An Internet trial on the illusory truth effect in schizophrenia and nonclinical participants. J. Behav. Ther. Exp. Psychiatry 43, 1058–1063 (2012).
Moritz, S., Andreou, C., Klingberg, S., Thoering, T. & Peters, M. J. V. Assessment of subjective cognitive and emotional effects of antipsychotic drugs. Effect by defect? Neuropharmacology 72, 179–186 (2013).
Moritz, S., Spirandelli, K., Happach, I., Lion, D. & Berna, F. Dysfunction by disclosure? Stereotype threat as a source of secondary neurocognitive malperformance in obsessive-compulsive disorder. J. Int. Neuropsychol. Soc. 24, 584–592 (2018).
Parrish, E. M. et al. Remote ecological momentary testing of learning and memory in adults with serious mental illness. Schizophr. Bull 47, 740–750 (2021).
Preiss, M., Shatil, E., Cermakova, R., Cimermannova, D. & Flesher, I. Personalized cognitive training in unipolar and bipolar disorder: a study of cognitive functioning. Front. Hum. Neurosci. 7, 108 (2013).
Rebchuk, A. D. et al. Assessment of prorated scoring of an abbreviated protocol for the National Institutes of Health Toolbox Cognition Battery. J. Int. Neuropsychol. Soc. 26, 1045–1050 (2020).
Siddi, S. et al. Comparison of the touch-screen and traditional versions of the Corsi block-tapping test in patients with psychosis and healthy controls. BMC Psychiatry 20, 1–10 (2020).
Stain, H. J. et al. The feasibility of videoconferencing for neuropsychological assessments of rural youth experiencing early psychosis. J. Telemed. Telecare 17, 328–331 (2011).
The authors would like to thank Philippine Hollander, Marianne Khalil, and Vanessa McGrory for their invaluable help during article selection and data extraction. This research was funded by a knowledge synthesis grant from the Canadian Institutes of Health Research COVID-19 rapid research funding opportunity in mental health and substance abuse (FRN: 171710). Initial reports are available online at: https://cihr-irsc.gc.ca/e/52066.html.
Dr. Lepage reports grants and personal fees from Otsuka/Lundbeck Alliance, grants from diaMentis, personal fees from Janssen, personal fees from MedAvante prophase, a grant from RBC Foundation, outside the submitted work. Dr. Lepage’s lab has also benefited from the Cogstate academic research support program which provides access to their cognitive assessment platform.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Lavigne, K.M., Sauvé, G., Raucher-Chéné, D. et al. Remote cognitive assessment in severe mental illness: a scoping review. Schizophr 8, 14 (2022). https://doi.org/10.1038/s41537-022-00219-x