# paper.pdf ## Metadata - PDFFormatVersion=1.6 - Language=en-US - IsLinearized=false - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - Author=Aziz U. R. Asghar - CreationDate=D:20260719084258+05'30' - Creator=Adobe InDesign 19.5 (Windows) - ModDate=D:20260719084318+05'30' - Producer=Adobe PDF Library 17.0; modified using iTextSharp™ 5.5.13.3 ©2000-2022 iText Group NV (AGPL-version) - Subject=Asghar AUR, Yuen HK, Aksoy M, Salawu A and Baseler HA (2026) Long COVID affects working memory: assessment using a single rapid online test. Front. Hum. Neurosci. 20:1881458. doi: 10.3389/fnhum.2026.1881458 - Title=Long COVID affects working memory: assessment using a single rapid online test - Trapped=False - dc:title=Long COVID affects working memory: assessment using a single rapid online test - dc:creator=Aziz U. R. Asghar,Hong Kiu Yuen,Murat Aksoy,Abayomi Salawu,Heidi A. Baseler - dc:description=Asghar AUR, Yuen HK, Aksoy M, Salawu A and Baseler HA (2026) Long COVID affects working memory: assessment using a single rapid online test. Front. Hum. Neurosci. 20:1881458. doi: 10.3389/fnhum.2026.1881458 - dc:subject=cognitive dysfunction,COVID-19,long COVID,mental fatigue,post-acute COVID-19 syndrome,short-term memory - dc:rights=Copyright © 2026 Asghar, Yuen, Aksoy, Salawu and Baseler. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. ## Contents ### Page 1 Aziz U. R. Asghar1,2,3, Hong Kiu Yuen2,4, Murat Aksoy1,2,Abayomi Salawu2,4 and Heidi A. Baseler1,3,5* 1Department of Psychology, University of York, York, United Kingdom,2Hull York Medical School,University of Hull, Hull, United Kingdom,3York Biomedical Research Institute, University of York, York,United Kingdom,4Disability Medicine and Rehabilitation, Hull University Teaching Hospitals NationalHealth Service (NHS) Trust, Hull, United Kingdom,5Hull York Medical School, University of York, York,United Kingdom Objective: Long COVID, or post-COVID-19 syndrome, is characterized by persis-tent symptoms following SARS-CoV-2 infection, including cognitive impairmentssuch as “brain fog” that adversely affect quality of life. The aim of this study wasto evaluate the impact of long COVID on working memory using a single, rapid,anonymous online survey and visual working memory quiz. Methods: We analyzed working memory scores in relation to reported longCOVID status (clinically diagnosed, self-reported, and non-long COVID), age,number of COVID-19 infections, long COVID duration, and subjective ratings ofbrain fog severity and the overall life impact of symptoms. The study utilized theRapid Objective Working Memory Assessment (ROWMA), a brief, gamified visualrecognition task designed to maximize compliance and minimize fatigue. Results: A total of 1,064 participants aged 16 to over 85 years were recruited,with 39% reporting long COVID. Categorical regression revealed that long COVIDstatus was the strongest predictor of working memory performance, followedby age and number of infections. Participants with long COVID performed sig-nificantly worse on the working memory quiz than non-long COVID controls,with the lowest scores observed in the clinically diagnosed group. Furthermore,working memory scores declined with age, particularly among those aged 35 andolder, and decreased with multiple COVID-19 infections within the diagnosedgroup. Individuals in the diagnosed group reported the most severe brain fog andthe greatest life impact, both of which strongly correlated with lower memoryscores. Although longer long COVID duration was associated with lower memoryperformance, an exploratory analysis indicated this trend may have been influ-enced by the SARS-CoV-2 variant. Conclusion: Our results support the hypothesis that long COVID is associatedwith objective working memory impairment, particularly among individuals witha clinical diagnosis. These findings add to the growing body of evidence linkinglong COVID to objective cognitive difficulties. Moreover, the study advocates theuse of rapid, single-task online cognitive assessments as a practical, scalable, andlow-burden approach for evaluating working memory function in this population. KEYWORDS cognitive dysfunction, COVID-19, long COVID, mental fatigue, post-acute COVID-19syndrome, short-term memory OPEN ACCESS EDITED BY César Fernández-de-las-Peñas,Rey Juan Carlos University, Spain REVIEWED BY Fatih Bal,Sakarya University, Türkiye Somayeh Pour Mohammadi,Universidad Rey Juan Carlos, Spain *CORRESPONDENCE Heidi A. Baseler ![email logo]() RECEIVED 14 May 2026 REVISED 30 June 2026 ACCEPTED 06 July 2026 PUBLISHED 22 July 2026 CITATION Asghar AUR, Yuen HK, Aksoy M,Salawu A and Baseler HA (2026) LongCOVID affects working memory:assessment using a single rapid onlinetest. Front. Hum. Neurosci. 20:1881458. COPYRIGHT © 2026 Asghar, Yuen, Aksoy, Salawu andBaseler. This is an open-access articledistributed under the terms of the. The use, distribution orreproduction in other forums ispermitted, provided the originalauthor(s) and the copyright owner(s) arecredited and that the original publicationin this journal is cited, in accordancewith accepted academic practice. Nouse, distribution or reproduction ispermitted which does not comply withthese terms. [TYPE](https://www.frontiersin.org/journals/human-neuroscience#editorial-board) [PUBLISHED](https://www.frontiersin.org/journals/human-neuroscience#editorial-board) [DOI](https://doi.org/10.3389/fnhum.2026.1881458) ### Page 2 Introduction Long COVID, also known as post COVID-19 syndrome/condi-tion, broadly refers to the signs and symptoms that develop during orafter an infection with the SARS-CoV-2 virus that persists over anumber of months and has no alternative explanation (;;). A systematic review of long COVID preva-lence, which included 194 studies with 735,006 participants, foundthat 45% of those who survived COVID-19 continued to experiencesymptoms at approximately 4 months (). Approximately 1.9 million people in the United Kingdom self-reported long COVID symptoms with 41% experiencing symptomsfor at least 2 years and 79% reported that their symptoms negativelyimpacted their daily activities ();over 200 long COVID symptoms have been documented, and themost commonly reported were fatigue (72%), difficulty concentrating(51%), muscle aches (49%) and shortness of breath (48%). Substantial cognitive performance deficits in tasks requiringmemory, attention and reasoning have been reported for people withCOVID-19 symptoms that persisted for up to 2 years (;;;;;;). The cognitive impact of long COVID, often generally describedas “brain fog” (including problems with memory and performing dailytasks), is a major contributor to poor functioning and reduced qualityof life experienced by patients with long COVID (;;;). Arecent systematic review and meta-analysis of studies evaluating brainfog, cognitive function and mental health symptoms in people withlong COVID found that the average overall prevalence of memory losswas 21% and as high as 61% (). This consid-erable variation in memory loss prevalence likely reflects differencesin how memory was defined, assessed or self-reported across studies. Working memory is the process of temporarily storing andretrieving information relevant to performing the current task(;), and is integral ineveryday tasks including problem solving, holding a conversation,reading comprehension and decision making. Working memory is acore cognitive function that supports a wide range of higher-ordercognitive processes, including attention, executive functioning, rea-soning, learning, and comprehension (;). Consequently, impairment of working memory may have wide-spread effects on everyday cognitive functioning. In line with thisview, working memory deficits have been consistently reported whentesting cognitive function in long COVID (;;,). Therefore, investigatingworking memory provides a focused means of assessing a cognitiveprocess that is both fundamental to higher-order cognition and com-monly affected in long COVID. The mechanisms underlying cognitive dysfunction in longCOVID remain incompletely understood. However, converging evi-dence suggests that SARS-CoV-2 infection may have persistent effectson brain systems that support attention, executive function, andmemory (;). Workingmemory relies on the coordinated activity of distributed neural net-works, particularly frontoparietal regions involved in the temporarystorage and manipulation of information (;). Disruption of these networks, whether through inflammatory,vascular, or other post-viral processes, may contribute to the memorydifficulties frequently reported by individuals with long COVID. Thehigh prevalence of self-reported cognitive symptoms, together withaccumulating evidence of objective cognitive deficits, highlights theneed for sensitive measures capable of detecting working memoryimpairment in this population. Studies investigating working memory deficits related to COVID-19 or long COVID are often measured within the context of a batteryof multiple cognitive tests (;;;;;,;). This is potentially problematic for tworeasons. Firstly, when embedded within a more complex cognitivetask (for example, spatial mental rotation), it is difficult to interpretperformance on the working memory component alone. Moreover, itis possible that a participant’s performance on a working memory taskcould be affected by previous different cognitive tasks. Secondly,undertaking numerous cognitive tests within a study requires correc-tion for multiple comparisons, which reduces statistical power. Sincemost people with long COVID report fatigue and difficulty concen-trating, completing a lengthy survey with multiple cognitive testswould be burdensome and may decrease completion rates or yieldunreliable data. Using the Rapid Objective Working MemoryAssessment (ROWMA), we previously demonstrated that COVID-19infection is associated with significantly lower working memoryscores compared to no prior infection (). The main aim of our study was to investigate the impact of longCOVID on working memory function. To capture a broad demo-graphic sample, we distributed an online survey with a single workingmemory quiz that could be completed relatively rapidly. We used thesame anonymous memory quiz previously shown to detect the nega-tive effects of COVID-19 on working memory. The quiz incorporatedgamified elements to encourage greater engagement and increasecompletion rates (). In our analysis we evaluate theimpact of several variables on objective memory scores including longCOVID status (clinically diagnosed long COVID, self-reported longCOVID and non-long COVID), age, gender, number of COVID-19infections and long COVID duration. We also examined the relation-ship between objective memory scores and subjective ratings of brainfog, and overall impact of long COVID symptoms on their life. Based on previous evidence of cognitive difficulties associatedwith long COVID, we hypothesized that individuals with long COVIDwould demonstrate poorer working memory performance than thosewithout long COVID. We further hypothesized that poorer workingmemory performance would be associated with increasing age, ahigher number of COVID-19 infections, and longer duration of longCOVID symptoms. Finally, we hypothesized that participants report-ing greater levels of brain fog and a greater overall impact of longCOVID symptoms on their daily lives would exhibit lower objectivememory scores. Materials and methods Ethics and participant recruitment The study was undertaken in accordance with the Declarationof Helsinki (). Ethical approvalwas given by the Hull York Medical School Ethics Committee(Reference 22–23 41). The survey and Rapid Objective Working ### Page 3 Memory Assessment (ROWMA) were made publicly available anddata collected from 16/03/2023 to 31/12/2023. Participants withand without long COVID (aged 16+) were recruited via profes-sional networks, social media, long COVID support groups andcommunity advertisements. Participants provided writteninformed consent via a mandatory “Yes/No” prompt before start-ing the study. The survey and ROWMA memory quiz were fullyanonymous, GDPR-compliant, and collected no personal contactdetails. Encrypted data was stored on the University of York’sQualtrics platform with access restricted to the investigators. Nocompensation was provided. Previous studies have inferred a diagnosis of long COVID basedon ongoing symptoms (;;;). In our study, we adoptedthe NHS England definition post-COVID- syndrome (long COVID),“…signs and symptoms that develop during or after COVID-19 andcontinue for more than 12 weeks and are not explained by an alternativediagnosis” (), which is broadly consistent with that of. To mitigate selection bias and account for healthcare barriers(e.g., lack of access or clinical validation), our methodology includesboth medically diagnosed and self-reported cases as well as thosewithout long COVID. This approach ensures a representative samplewhile allowing for a direct comparison between the three groups. Participants were classified as having diagnosed long COVID ifthey reported receiving a diagnosis of long COVID from a healthcareprofessional. In the United Kingdom, long COVID is recognized as acondition in which symptoms develop during or after an infectionconsistent with COVID-19, persist for more than 12 weeks, andcannot be explained by an alternative diagnosis (). Clinicaldiagnosis is typically based on an assessment of symptoms and dura-tion, and may be supported by further investigations or referral tospecialist post-COVID services where appropriate. Participants whoreported persistent symptoms consistent with long COVID but hadnot received a formal diagnosis from a healthcare professional wereclassified as the self-reported long COVID group. Long COVID survey and memory quizstructure The online assessment comprised a survey component followedimmediately by an objective working memory quiz, structured simi-larly to our previous online evaluation paradigm ().Both elements were hosted via Qualtrics (University of York) andcould be accessed on any internet-enabled device using a URL or QRcode. Screenshots of example survey questions and a sample of thememory quiz are illustrated in. The survey component consisted of a series of “Yes/No” questionsasking participants about their long COVID status. First, we asked,“Have you been diagnosed with long COVID (post-COVID-19 syn-drome) by a medical professional?.” If a participant answered “Yes,”they were designated into the diagnosed long COVID group. Thoseanswering “No” were then asked, “Do you think that you may havelong COVID?,” and we included the NHS England definition of longCOVID (post COVID-19 syndrome): “Signs and symptoms thatdevelop during or after COVID-19 and continue for more than12 weeks and are not explained by an alternative diagnosis” (). Those that answered “Yes” to this question were designated asthe self-reported long COVID group. Those who answered “No” toboth questions were designated as the non-long COVID group. If theywere diagnosed with long COVID or self-reported having longCOVID, they were asked how many months it had been since theirlong COVID symptoms started (using a drop-down menu, from lessthan 3–60 months). We next asked long COVID participants to sub-jectively rate the overall life impact of their long COVID symptoms(using “None,” “Mild,” “Moderate,” or “Severe”). All participants(regardless of long COVID status) were asked whether they had beeninfected with COVID-19 in the last 12 weeks (“Yes/No”), and howmany times they had been infected with COVID-19 (“Not hadCOVID-19,” “One,” “Two,” “Three,” or “Four or more times”). All par-ticipants were also asked to subjectively rate their level of “brain fog”(using “None,” “Mild,” “Moderate,” or “Severe”). The following ![Panel a shows screenshots from a Long COVID survey and memory quiz, including introductory questions and various memory tasks requiring participants to recall and select images or numbers they recently viewed. Panel b depicts the memory task procedure, showing a sequence where two animal pictures are presented sequentially, followed by a multiple-choice selection of previously seen images, with feedback on correctness. Arrows and labels indicate time progression and inter-stimulus intervals.]() FIGURE 1 The “Long COVID Survey and Memory Quiz” presented on a smartphone. (a) Example questions from the survey, and examples of the animal, number,fruit and object categories shown from the memory quiz. (b) A series of screenshots representing a single trial from the working memory quiz. ### Page 4 definition of “brain fog” was provided to aid participants: “for exam-ple, difficulty in thinking clearly or quickly, concentrating, having aconversation, remembering things, or getting tired easily after mentaleffort.” This definition of brain fog was based on our previous findingsshowing that these COVID-19 cognitive symptoms were highly cor-related with one another (). Finally, participantswere asked to select their age range (16–17, 18–24, 25–34, 35–44,45–54, 55–64, 65–74, 75–84, and 85+), gender, and country of currentresidence. Following the completion of the survey, participants undertookthe working memory quiz. To ensure data quality, enhance compli-ance, and maximize concentration, participants were instructed tocomplete this subsequent task in a quiet, distraction-free environ-ment. This objective working memory assessment was developedusing established principles of cognitive testing and employed a visualrecognition paradigm to evaluate the short-term storage and retrievalof visual information. The task was adapted from widely used visualpicture span and digit span paradigms, with stimuli presented acrossmultiple semantic categories (;). The construct validity of this assessment is supportedby its application in a previous study using the same online visualworking memory task, where it successfully discriminated betweenindividuals with and without COVID-19 infection, demonstrated sen-sitivity to infection severity (hospitalized versus non-hospitalized),and significantly correlated with participants’ self-reported cognitivefunction and symptom severity (). Moreover, ourworking memory quiz, known as the Rapid Objective WorkingMemory Assessment (ROWMA), has been incorporated into a pub-lished randomized controlled pilot trial protocol approved by theHealth Research Authority England (IRAS: 312024), Patient-LedCognitive Gamified Training During Hemodialysis. In that study,ROWMA was administered alongside established cognitive assess-ments, including the Montreal Cognitive Assessment (MoCA) andModified Mini-Mental State Examination (3MSE), allowing compari-son with widely used measures of cognitive function (). The working memory quiz was intentionally designed to bebrief and engaging to maximize completion rates and minimizeparticipant fatigue, which is particularly important when testingindividuals with long COVID who frequently report fatigue as aprimary symptom. The memory task’s design prioritizes ecologicalvalidity and accessibility over the laboratory-controlled conditionstypical of traditional cognitive assessments, representing a trade-off that allowed us to collect data from a larger and more diversesample. To increase engagement, the ROWMA task employedaspects of gamification, ensuring the protocol was easy to under-stand through the use of familiar and engaging images. The com-plete stimulus set comprised 55 unique images, including 15unique images of animals, 15 unique images of fruits, 15 uniqueimages of objects, and 10 single-digit whole numbers rangingfrom 0 to 9. During the task, participants progressed throughincreasing difficulty levels, requiring them to recall sequences oftwo to five random images shown consecutively across these foursemantic categories. Each stimulus image was displayed for aduration of 500 ms, with an interstimulus interval randomlyvaried between 500 and 600 ms. Following the sequence presenta-tion, a recognition grid of six images was displayed, and partici-pants were instructed to select the precise target items they hadjust viewed. Participant responses and overall task completionwere entirely untimed and free from any time restrictions.Throughout the assessment, dynamic motivational feedback wasprovided following participants’ responses to maintain engage-ment. Upon completion of the entire quiz, participants were auto-matically notified of their overall performance score along withtheir individual scores for each of the four specific semanticsub-categories. Data analysis We only included data from participants who had completed thesurvey and memory quiz in its entirety. Based on our inclusion crite-ria, participants under the age of 16 were excluded (N = 3).Participants that had COVID-19 infections within the last 12 weekswere also excluded from the analysis (N = 146), as a recent COVID-19infection is known to affect working memory (). Inaddition, we excluded one participant who reported having longCOVID for 49 months before completing the survey/memory quiz on21st March 2023. This meant that their long-COVID symptoms wouldhave started on the 21st of February 2019, a time before COVID-19was identified. A score of 1 was given for each correct trial in the memory quiz,yielding a possible maximum objective memory score of 16. StatisticalPackage for the Social Sciences (SPSS Version 29.0.1.0, IBMCorporation, Armonk, N. Y., United States) was utilized to performstatistical analyses. To minimize the risk of Type I errors from multiplecomparisons while maintaining sensitivity to potential effects, weemployed a hierarchical analytical approach. Categorical regressionwith optimal scaling (CATREG; SPSS) was selected because our studyincluded a combination of nominal, ordinal, and continuous variables.CATREG uses optimal scaling to assign numerical values to categori-cal variables in a manner consistent with their measurement level,enabling variables measured on different scales to be incorporatedinto a single multivariable regression model. We used categoricalregression to identify which independent variables significantly con-tributed to variance in memory scores. In the first categorical regres-sion, we assessed the contributions of the following independentvariables: long COVID status (three categorical/nominal groups: non-long COVID, diagnosed long COVID and self-reported long COVID),age (nine ordinal groups: 16–17, 18–24, 25–34, 35–44, 45–54, 55–64,65–74, 75–84, and 85+), number of COVID-19 infections (five ordinalgroups: none, 1, 2, 3, and 4 or more) and gender (three categorical/nominal groups: female, male, other/prefer not to say). A second cat-egorical regression was performed using the long COVID groups onlywith the following independent variables: long COVID status (twogroups: diagnosed long COVID and self-reported long COVID), age,number of COVID-19 infections, gender, and long COVID durationin months (ordinal, from less than 3 months up to 60 months). Next,we investigated the effect of the significant individual variables arisingfrom the categorical regression analysis on objective memory scores,using the following non-parametric statistics, as the data were non-normally distributed: Mann–Whitney U test and Kruskal-Wallis testwith Dunn’s pairwise comparisons (Bonferroni corrected). We evalu-ated relationships between pairs of variables using the Kendall’s tau-bcorrelation coefficients, as this is the most appropriate metric in cor-relation analyses involving ordinal variables with tied ranks. All test-ing was two-tailed, and we set statistical significance levels to p-valuesless than 0.05. ### Page 5 Results Participant demographics Participant demographics (N = 1,064) are detailed in.Only those who fully completed the survey and memory quiz wereincluded. Over 94% of participants (N = 1,001) completed the surveyand memory quiz within 15 min (median = 7.2 min): long COVIDgroup (median = 7.3 min, N = 393); non-long COVID group(median = 6.9 min, N = 608). Categorical regression and objectivememory scores A categorical regression revealed that significant effects werefound, in order of importance, for long COVID status, age andnumber of COVID-19 infections (). Gender did not have asignificant effect on memory scores. A second categorical regressionwas performed for the two long COVID groups only to assess theadditional contribution of long COVID duration on memory scores. shows that the significant factors were, in order of impor-tance: age, long COVID duration, and number of COVID-19 infec-tions. Long COVID subgroup or gender did not have a significanteffect on memory scores. We then went on to evaluate each of thesignificant factors independently. We used non-parametric statisticsin all these subsequent analyses as a Kolmogorov–Smirnov testrevealed that objective memory scores were not normally distributed[D(1,064) = 0.26, p < 0.001]. Effect of long COVID status on workingmemory shows that the objective memory scores are signifi-cantly reduced in the two long COVID groups combined [Mean(M) ± standard error of the mean = 14.15 ± 0.10] compared to thenon-long COVID group (M = 15.02 ± 0.05), Mann–WhitneyU = 167938.50, z = −7.12, p < 0.001, r = 0.22. Next, we divided thelong COVID group into subgroups: those who suspected they mayhave long COVID (self-reported), and those who were diagnosedwith long COVID by a medical professional. shows thatobjective memory scores were highest in the non-long COVID groupbut were significantly reduced in each of the long COVID subgroups,with the lowest scores in the diagnosed group. A Kruskal-Wallis testrevealed a significant difference between the mean memory scoresacross groups, H(2) = 61.38, p < 0.001, η2 = 0.06. Dunn’s pairwisecomparison tests with Bonferroni correction revealed significant TABLE 1 Participant demographics. Data values are presented as sample size (%). ### Page 6 differences between the non-long COVID group (M = 15.02 ± 0.05)and the self-reported long COVID (M = 14.48 ± 0.14), z = 3.20,p = 0.0042, r = 0.11, and between the non-long COVID and diag-nosed long COVID groups (M = 13.88 ± 0.15), z = −7.71, p < 0.001,r = 0.26. In addition, there was a significant difference between theself-reported and diagnosed long COVID groups, z = −3.28,p = 0.0031, r = 0.16. Effect of age on working memory We found that objective memory scores decreased significantlywith age for both the non-long COVID and the diagnosed longCOVID groups but not for the self-reported long COVID group. AKendall’s tau-b correlation showed a negative relationship betweenmean memory score and age for both the non-long COVID(τb = −0.155, p < 0.001) and diagnosed long COVID (τb = −0.208,p < 0.001) groups across the nine age categories. The self-reportedlong COVID group did not have a significant relationship betweenage and mean memory scores (τb = −0.099, p = 0.098). Participantsaged 16–17 years old were combined with those aged 18–24 yearsold, and participants above 55 years old were combined to increasecategory sample size (see). shows the effect of ageon memory scores for all three groups for the five age categories. AKruskal-Wallis test was performed at each of the five age categoriesto determine differences between groups based on long COVIDstatus. There was no significant effect of long COVID status onmemory scores for the 16–24 year-old age category [H(2) = 4.201,p = 0.122, η2 = 0.02], or the 25–34 year-old age category[H(2) = 0.004, p = 0.998, η2 = 0]. However, there was a significanteffect of long COVID status on memory scores for the 35–44[H(2) = 18.539, p < 0.001, η2 = 0.08], 45–54 [H(2) = 24.377,p < 0.001, η2 = 0.09] and 55 + year-old categories [H(2) = 15.416,p < 0.001, η2 = 0.05]. Dunn’s pairwise comparisons (withBonferroni correction for multiple tests) revealed that the memoryscores in the diagnosed long COVID group were significantlyreduced compared to the non-long COVID group (all p < 0.001) inthe 35–44 (z = −4.01; r = 0.30), 45–54 (z = −4.88; r = 0.33), and55 + year-old categories (z = −3.88; r = 0.27). The memory scoresfor the self-reported long COVID group were significantly reducedcompared to the non-long COVID group in the 35–44 age category(z = −2.70; r = 0.21; p = 0.021). In the 55 + age category, thememory scores for the diagnosed long COVID group were signifi-cantly smaller than for the self-reported group (z = −2.53; r = 0.24;p = 0.034). Effect of the number of COVID-19infections on working memory Objective memory scores decreased significantly with increasingnumber of COVID-19 infections for the diagnosed long COVID group(τb = −0.113, p = 0.043). However, there was no significant correlationbetween the number of infections and memory scores for either the self-reported long COVID group (τb = −0.040, p = 0.527) or the non-longCOVID group (τb = 0.001, p = 0.970). Next, we compared the impact ofthe number of COVID-19 infections on memory scores between thenon-long COVID, self-reported and diagnosed long COVID groups.Since there were only 13 participants with four or more COVID-19infections we combined these participants with those who had threeinfections (N = 67). shows the effect of the number of COVID-19 infections on memory scores for all three long COVID status groupsfor the four infection categories. A Kruskal-Wallis test was performed ateach of the four infection categories to determine differences betweengroups based on long COVID status. When participants had one or twoCOVID-19 infections more than 12 weeks prior to completing the TABLE 2 Categorical regression analysis and objective memory scores for non-long COVID, diagnosed long COVID and self-reported long COVID participants. Significant p-values are given in bold. TABLE 3 Categorical regression analysis and objective memory scores for diagnosed long COVID and self-reported long COVID participants only. Significant p-values are given in bold. ### Page 7 survey and memory quiz, there was a significant effect of long COVIDstatus on memory scores [one COVID-19 infection: H(2) = 23.046,p < 0.001, η2 = 0.04; two COVID-19 infections: H(2) = 25.352, p < 0.001,η2 = 0.09]. There was no significant effect of long COVID status onmemory scores in participants with 3 or more infections [H(2) = 5.228,p = 0.073, η2 = 0.04]. Dunn’s pairwise comparisons (with Bonferroni cor-rection for multiple tests) revealed that the memory scores in the diag-nosed long COVID group were significantly reduced compared to thenon-long COVID group (p < 0.001) in the one (z = −4.70; r = 0.23) andtwo COVID-19 infection categories (z = −5.00; r = 0.34). The memoryscores for the diagnosed long COVID group were significantly reducedcompared to the self-reported group in the two-infection category(z = −3.05; r = 0.25, p = 0.007). For the zero-infection category, there wasan insufficient number of participants to perform statistical analysis(these participants stated having long COVID despite an apparentabsence of COVID-19 infection). Effect of long COVID duration on workingmemory As the number of months of having long COVID increased, thememory scores decreased significantly in the diagnosed long COVIDgroup (τb = −0.178, p < 0.001) but not in the self-reported longCOVID group (τb = −0.021, p = 0.712). Next, we compared theimpact of long COVID-19 duration on memory scores between theself-reported long COVID and diagnosed long COVID groups. Wecombined participants into six-month long COVID-19 duration cat-egories. shows the effect of long COVID duration on memoryscores for the self-reported long COVID and diagnosed long COVIDgroups. A Mann–Whitney U test found no significant differencebetween the self-reported long COVID and diagnosed long COVIDgroups for any of the five long COVID duration categories,Bonferroni-corrected for five duration categories. ![Bar chart illustrating mean memory scores across age groups from 16-24 to 55+, comparing Non-Long COVID, Self-Reported Long COVID, and Diagnosed Long COVID. Scores decline more steeply with age for Long COVID groups, with significant differences marked by asterisks. Error bars indicate variability, and sample sizes are shown below each bar.]() FIGURE 3 The effect of age on objective memory scores in the non-long COVID, self-reported long COVID and diagnosed long COVID groups. The sample sizeis given in brackets for each data bar. Error bars represent the standard error of the mean. Asterisks indicate significant differences between groups foreach age range category. *p < 0.05, **p < 0.001. ![Panel a\) shows a bar graph comparing mean memory scores between non-long COVID participants, who scored higher, and long COVID participants, with significant difference indicated by double asterisks. Panel b\) presents mean memory scores for non-long COVID, self-reported long COVID, and diagnosed long COVID groups, with a progressive decrease in scores and statistically significant differences between all groups denoted by single and double asterisks.]() FIGURE 2 The effect of long COVID status on objective memory scores. (a) Non-long COVID group compared to both long COVID groups combined (diagnosedplus self-reported long COVID), and (b) non-long COVID group compared to self-reported and diagnosed long COVID groups separately. Error barsrepresent standard error of the mean. *p < 0.01, **p < 0.001. ### Page 8 COVID-19 variants and working memory The SARS-CoV-2 virus evolved into different variants throughoutthe pandemic, meaning participants likely contracted different strainsdepending on when they were infected. Research indicates that thoseinfected during earlier variant waves face a higher probability ofpersistent long COVID symptoms, including cognitive and memoryimpairments (). Hence, lower memory scores asso-ciated with longer COVID duration () may have been influ-enced by the specific SARS-CoV-2 variant. In a post-hoc analysis,participants based in the United Kingdom were grouped by the vari-ant dominant at the onset of their long COVID symptoms. Adopting ![Bar chart illustrating mean memory scores by Long COVID duration in months, comparing self-reported and diagnosed Long COVID. Memory scores decrease as duration increases, with diagnosed cases showing a lower mean score, especially after nineteen months. Error bars indicate variability.]() FIGURE 5 The effect of long COVID duration on objective memory scores in the self-reported and diagnosed long COVID groups. The sample size is given inbrackets for each data bar. Each bar represents a different sample of participants (between-subject design). Error bars represent the standard error ofthe mean. ![Bar chart showing mean memory score by number of COVID-19 infections and group: non-long COVID, self-reported long COVID, and diagnosed long COVID. Memory scores are lower in long COVID groups, with statistically significant differences indicated for groups with one and two infections. Error bars and sample sizes are included for each bar.]() FIGURE 4 The effect of COVID-19 infections on objective memory scores in the non-long COVID, self-reported long COVID and diagnosed long COVID groups.The sample size is given in brackets for each data bar. Error bars represent the standard error of the mean. Asterisks indicate significant differencesbetween groups for the number of COVID-19 infections. The dashed border on vertical bars (zero COVID-19 infections category) indicates the smallnumber of participants who stated having long COVID despite an apparent absence of COVID-19 infection. *p < 0.01, **p < 0.001. ### Page 9 the methodology employed in major community-based epidemiologi-cal trials, the REal-time Assessment of Community Transmission(REACT) studies, we used the date of infection as a proxy for thedominant circulating strain to classify variant cohorts without requir-ing universal viral sequencing, which is often impractical andresource-intensive in large population-based studies (;). This onset was estimated by subtracting thereported symptom duration (months) from the date the survey andquiz were completed. Our variant groups were based on the time peri-ods used previously for the United Kingdom: (1) wild-type, prior toDecember 2020; (2) alpha, from the beginning of December 2020 tothe end of April 2021; (3) delta, from the beginning of May 2021 to theend of December 2021; (4) omicron, from the beginning of January2022 onwards (;). A Kendall’s tau-b correlation showed a significant relationshipbetween the SARS-CoV-2 variant that was dominant at the time atwhich long COVID symptoms started and the duration of long COVIDsymptoms for both the self-reported (τb = −0.740, p < 0.001) and diag-nosed (τb = −0.869, p < 0.001) long COVID groups. illustratesthe objective memory scores for each dominant SARS-CoV-2 variantin the self-reported and diagnosed long COVID groups in UnitedKingdom participants only. For participants with diagnosed longCOVID, there was a significant difference in mean memory scoresacross the four SARS-CoV-2 variants [Kruskal-Wallis test,H(3) = 14.140, p = 0.003, η2 = 0.07]. For participants with diagnosedlong COVID, we found significantly lower mean memory scores for thewild type compared to the omicron variant (z = 2.78; r = 0.27,p = 0.033) and alpha compared to the omicron variant (z = −3.17;r = 0.34, p = 0.009). There was no significant difference in meanmemory scores across the four SARS-CoV-2 variants for participantswith self-reported long COVID [Kruskal-Wallis test, H(3) = 1.236,p = 0.744, η2 = 0.01]. Relationship between “brain fog,” “impactof long COVID symptoms on life” andobjective memory scores There was a significant negative correlation between the subjec-tive ratings for “Brain fog” and objective memory scores across allparticipants (τb = −0.185, p < 0.001). shows the mean sub-jective ratings for “Brain fog” versus objective memory scores for eachof the non-long COVID, self-reported long COVID and diagnosedlong COVID groups. A Kruskal-Wallis test revealed a significant rela-tionship between long COVID status and “Brain fog” ratings[H(2) = 331.90, p < 0.001, η2 = 0.31]. Dunn’s pairwise comparisons(with Bonferroni correction for multiple tests) revealed that com-pared to the non-long COVID group, subjective ratings of “Brain fog”were significantly more severe in the self-reported (z = −11.61;r = 0.40, p < 0.001) and diagnosed long COVID groups (z = 16.42;r = 0.55, p < 0.001). Moreover, the subjective ratings of “Brain fog” inthe diagnosed long COVID group were significantly more severecompared to the self-reported long COVID group (z = 2.92; r = 0.14,p = 0.011). There was a significant negative correlation between thesubjective ratings for “Impact of long COVID symptoms on life” andobjective memory scores across all participants with long COVID(τb = −0.169, p < 0.001). shows the mean subjective ratingsof the “Impact of long COVID symptoms on life” versus objectivememory scores for the self-reported and diagnosed long COVIDgroups. The subjective ratings of “Impact of long COVID symptoms onlife” in the diagnosed long COVID group were significantly worsecompared to the self-reported long COVID group (Mann–WhitneyU test = 10372.00, z = −9.61, p < 0.001, r = 0.47). There was a signifi-cant correlation between the subjective ratings for “Brain fog” and“Impact of long COVID symptoms on life” in the long COVID (self-reported and diagnosed) group (τb = 0.405, p < 0.001). ![Bar chart comparing mean memory scores for self-reported and diagnosed long COVID across four SARS-CoV-2 variants: wild type, Alpha, Delta, and Omicron; shows increasing scores from wild type to Omicron with statistical significance between groups indicated by asterisks. Error bars represent standard deviations and sample sizes are noted for each group.]() FIGURE 6 The dominant SARS-CoV-2 variant and objective memory scores in the self-reported and diagnosed long COVID groups in participants only from theUnited Kingdom. The long COVID participants were divided into four groups (Wild type, Alpha, Delta and Omicron) according to the dominant variant atthe time when their long COVID symptoms began. The sample size is given in brackets for each data bar. Error bars represent the standard error of themean. *p < 0.05, **p < 0.01. ### Page 10 Discussion Our study supports the hypothesis that long COVID may affectworking memory. Long COVID status (diagnosed or self-reported)was the strongest predictor among the variables examined, followedby age and infection frequency. Using the same rapid online test as, we confirmed that both long COVID groups per-formed significantly worse than non-long COVID controls. Theseresults align with previous studies reporting memory deficits in indi-viduals with continuing COVID-19 symptoms (;;;,;). Our diagnosed long COVID group’s mean score(13.88) was lower than non-hospitalized COVID-19 patients (14.73)and more comparable to hospitalized patients (13.46) from our previ-ous work (), suggesting poorer working memoryperformance in this population. Our results link working memory deficits to “brain fog,” a hall-mark of long COVID (;;;). Objective memory scores negativelycorrelated with brain fog severity, which was highest in the diagnosedgroup, followed by self-reported and non-long COVID groups. Thisfinding aligns with the significantly higher rates of memory impair-ment observed in long COVID participants reporting brain fog(). The diagnosed long COVID group reported a more severeimpact on daily life with lower memory scores than the self-reported group. These findings suggest that working memory defi-cits contribute to the overall functional burden of the condition, alink further supported by the strong correlation between memoryscores, subjective brain fog and the reported impact of symptomson life. This aligns with previous studies showing that long COVIDcan significantly reduce quality of life, including persistent physicalsymptoms, mental health problems, ability to cope, reduced func-tional status and employment issues (;;;;;). Previous studies on cognitive function in long COVID have typi-cally classified participants into a long COVID group based on persis-tent symptomology rather than clinical diagnosis by a medicalprofessional (;;;). One of the strengths of our study is thatwe included a group of long COVID participants who were clinicallydiagnosed. Medical professionals will diagnose patients with longCOVID based on multiple sources of evidence, including case history,signs and symptoms, physical assessments/clinical tests, and may besupported by members of a multi-disciplinary team. In our study, weadded a separate self-reported long COVID group to represent thosewho may have long COVID but were not clinically diagnosed. Objective memory scores in the non-long COVID groupdecreased significantly with age, consistent with our previous study() and others (;). Notably, this age-related decline was more pronounced inthe diagnosed long COVID group. While memory scores were com-parable between diagnosed and control groups under age 35, workingmemory scores were significantly lower in the diagnosed long COVIDgroup for those aged 35 and above. Our finding that long COVID maybe an additional factor contributing to memory impairment in olderadults aligns with broader evidence that aging increases vulnerabilityto post-COVID-19 cognitive deficits (;;;;;;). SARS-CoV-2 reinfection increases the risk of persistent neuro-logical symptoms (). Consistently, we found thatmemory scores decreased as the number of infections increased, spe-cifically within the diagnosed long COVID group. This supports theview that reinfection exacerbates cognitive dysfunction in these indi-viduals. A small number of participants (N = 6) reported long COVIDdespite no known infection. This could be due to a lack of initial test-ing, asymptomatic infection (), vaccine side effects(), or reporting errors. In our cross-sectional study, memory scores decreased as longCOVID duration increased. However, this association may be ![Two-panel line graph compares mean memory scores among groups with varying levels of 'brain fog' \(panel a\) and 'impact of long COVID symptoms on life' \(panel b\). In both panels, three groups are represented: non-long COVID \(green square\), self-reported long COVID \(yellow circle\), and diagnosed long COVID \(red diamond\). In panel a, mean memory score decreases from non-long COVID to diagnosed long COVID, with statistical significance indicated for group differences. In panel b, a similar decline in score is observed, with fewer severity categories represented and significant differences noted between groups.]() FIGURE 7 Subjective ratings of (a) “Brain fog” and (b) “Impact of long COVID symptoms on life” versus memory scores. The vertical error bars represent thestandard error of the mean for memory scores, and horizontal error bars indicate the standard error of the mean for “Brain fog” or “Impact of longCOVID symptoms on life” ratings. Statistical significance values refer to group differences in ratings for “Brain fog,” and for “Impact of long COVIDsymptoms on life” *p < 0.01, **p < 0.001. ### Page 11 influenced by differences in the SARS-CoV-2 variants circulatingat the time participants developed long COVID symptoms.Previous studies have reported that earlier variants may be associ-ated with a greater risk of persistent cognitive symptoms than latervariants (). In a post-hoc exploratory analy-sis, we divided participants from the United Kingdom with longCOVID into four variant groups (wild-type, alpha, delta and omi-cron) based on the dominant SARS-CoV-2 strain circulating duringthe estimated onset of their long COVID symptoms (;). Participants in the diagnosed longCOVID group whose symptoms were reported to have begunduring periods dominated by the wild-type and alpha variants hadlower memory scores than those assigned to the omicron period.However, because variant classification was based on estimatedperiods of variant predominance rather than confirmed sequencingdata, these findings should be interpreted with caution and consid-ered hypothesis-generating. Future studies using confirmed infec-tion dates and variant identification are needed to determine theextent to which SARS-CoV-2 variants contribute to long-term cog-nitive outcomes. Long COVID status was based on participant self-report ratherthan independent clinical verification, and some degree of misclas-sification cannot be excluded. Nevertheless, the diagnosed longCOVID group had lower memory scores than the self-reportedgroup, suggesting that working memory deficits were more severein those diagnosed by a medical professional. Furthermore,memory scores in the self-reported long COVID group showed nosignificant decline associated with age, infection frequency, longCOVID duration, or probable viral variant. Participants with moresevere symptoms, such as brain fog, memory problems, and greaterlife disruption, may have been more likely to seek medical help andreceive a formal long COVID diagnosis, whereas those with mildersymptoms may have been less inclined to consult a professional andtherefore had higher memory scores in the self-reported group.Our study supports this hypothesis, as we found that the diagnosedlong COVID group reported that they had more severe brain fogand that their symptoms had a greater impact on their life. A previ-ous study has shown that more severe COVID-19 symptoms arelinked to poorer cognitive function including memory deficits(). Several mechanisms have been proposed to explain the impact oflong COVID on cognition including memory function. These includepathophysiological mechanisms that could impede brain functionsuch as neuroinflammation, damage to neurons and astrocytes, vas-cular alterations, neurotropic effects, dysautonomia, hypoxia, bloodcoagulation, and metabolic changes (;;;;;;). In addition, long COVIDcould affect psychological wellbeing, leading to conditions such asanxiety, depression, post-traumatic stress disorder, or sleep distur-bances which in turn could contribute to memory deficits (;;;;). A number of rehabilitation approaches have been proposed totreat long COVID cognitive symptoms (with mixed effectiveness)including cognitive training, electromagnetic brain stimulation, psy-chosocial, pharmaceutical, natural supplements, psychoeducation/compensatory skills training, telerehabilitation and physical rehabili-tation such as hyperbaric oxygen therapy (;;;). A personal-ized clinical treatment plan that considers working memory deficitsmay help address both the cognitive as well as physical symptoms oflong COVID. Study limitations One limitation of our study is the use of a single online memoryquiz that has not undergone formal psychometric validation againstgold-standard neuropsychological assessments. While the quiz wasdesigned based on established working memory paradigms and haspreviously demonstrated sensitivity to COVID-related cognitive dif-ferences (), we cannot make definitive claims aboutits reliability, validity, or specificity. However, our approach offers apragmatic solution for screening larger populations for workingmemory difficulties in long COVID, which may then inform moredetailed clinical evaluation. Our working memory quiz is currentlybeing applied along with several standard neuropsychological testsincluding the Montreal Cognitive Assessment (MoCA) and theModified Mini-Mental State Exam (3MSE) in a study evaluating cog-nitive function in patients with kidney disease undergoing hemodialy-sis (). Future research should aim to validate thisbrief online measure against established clinical neuropsychologicaltests to determine its psychometric properties, including test–retestreliability, construct validity, discriminant validity, and sensitivity tochange over time. Prior to starting the memory quiz, the survey reminded partici-pants of the importance of the study and to undertake the quiz in anenvironment without distractions. However, we acknowledge theinherent limitations of unsupervised online testing, including variablelevels of distraction and compliance, which could affect performanceof any participant regardless of long COVID status. Nonetheless,online testing offers significant advantages in terms of accessibility,particularly important for reaching individuals with mobility limita-tions including those with long COVID. The ecological validity oftesting participants in more natural environments may, however,better reflect real-world cognitive functioning than laboratory-basedassessments. Our use of convenience sampling through social media, supportgroups, and professional networks may have introduced bias. Forexample, individuals engaged in long COVID support groups may bethose who experience more persistent or severe symptoms, whichcould lead to an overestimation of cognitive difficulties. However,there may have also been an underrepresentation of those too unwellto complete the study, which could contribute to an underestimationof effects in the most affected group. The majority of our sample wasbased in the United Kingdom (83%), which may limit the generaliz-ability of the findings worldwide. Capturing demographics and epide-miology of long COVID across countries is challenging becausestudies across the world often use different sampling methods, diag-nostic criteria, and methodologies, making it difficult to consistentlycompare and combine datasets. However, the distribution of ages andgender in our sample are in line with the demographics published ina comprehensive review of long COVID which includes data from theOffice of National Statistics in the United Kingdom (). Although an a priori power analysis was not conducted beforedata collection, with our final sample size of 1,064 participants (415with long COVID and 649 controls) and an effect size of Hedges’ ### Page 12 g = 0.552, the calculated statistical power was > 99%. Moreover, aftercorrection for multiple comparisons, all statistically significant differ-ences resulted in p < 0.005. Together, this indicates that our study issufficiently powered to detect differences in working memory scoresusing our single rapid online test. The purpose of the study was to compare working memoryfunction in those with long COVID with those without longCOVID, rather than to determine the specific contribution of allfactors that may influence cognitive performance. Although sev-eral variables were significantly associated with working memoryperformance, the regression models explained only a modest pro-portion of the variance in memory scores (R2 = 0.114 and 0.079),indicating that additional factors not assessed in the present studyare likely to contribute to individual differences in workingmemory function. We operated under the assumption that partici-pants with pre-existing cognitive impairments or other conditionsaffecting working memory were equally distributed across thenon-long COVID and long COVID cohorts, but this may not havebeen the case. Long COVID is a complex and heterogeneous con-dition that may disproportionately affect specific demographic andsocioeconomic groups (). Furthermore,fatigue, memory problems, sleep disturbance, anxiety, depression,and other co-occurring symptoms are common in individuals withlong COVID and may contribute to variability in cognitive out-comes (;;;;;;). Educational attainmentand socioeconomic status have also been associated with differ-ences in cognitive and executive functioning, including memoryand working memory performance (). As edu-cational attainment, socioeconomic status, medication use, andpre-existing neurological or psychiatric conditions were notassessed in the present study, their potential contribution to theobserved differences in working memory performance could notbe evaluated. Consequently, residual confounding cannot beexcluded, and the variables identified in this study should be inter-preted as significant correlates of working memory performancerather than comprehensive determinants of cognitive outcomes inlong COVID. To maximize engagement, our memory quiz was intentionallydesigned to be relatively easy. Consequently, while the observed dif-ferences were significant, effect sizes were small-to-medium, yield-ing mean score differences of just 1–2 points out of a possible 16.Nevertheless, even modest reductions in working memory mayaffect everyday activities such as problem-solving, decision-making,reading comprehension, and maintaining conversations. Consistentwith this, lower memory scores were associated with greater self-reported brain fog and a more severe impact of long COVID symp-toms on daily life. While the cross-sectional design does not permitcausal inference and the regression models explained only a modestproportion of the variance in memory scores, these findings suggestthat working memory impairment may represent a clinically rele-vant component of the cognitive difficulties experienced by someindividuals with long COVID. Consequently, the observed associa-tions should be viewed as contributing to the growing evidence thatworking memory may be affected in long COVID, while recognizingthat additional factors not assessed in the present study are alsolikely to influence cognitive outcomes. Our working memory quizmay have value as part of a broader assessment to help identify andcharacterize working memory difficulties in individuals withlong COVID. Conclusion Our results suggest that long COVID is associated with poorerworking memory performance, particularly among participantsreporting a clinical diagnosis of long COVID. Working memoryperformance was also associated with age, infection frequency,brain fog severity, and the perceived impact of symptoms on dailylife. While these findings are consistent with cognitive difficultiesreported in long COVID, they should be interpreted in the contextof the study’s cross-sectional design, online assessment methodol-ogy, and the possibility that additional factors not measured in thepresent study may have influenced working memory performance.Nevertheless, our findings add to the growing body of evidencelinking long COVID with objective cognitive impairment and sup-port the use of rapid online assessments as a practical approach forlarge-scale evaluation of working memory difficulties in thispopulation. Data availability statement The raw data supporting the conclusions of this article will bemade available by the authors, without undue reservation. Ethics statement The studies involving humans were approved by the Hull YorkMedical School Ethics Committee. The studies were conducted inaccordance with the local legislation and institutional requirements.The participants provided their written informed consent to partici-pate in this study. Author contributions AA: Project administration, Conceptualization, Formal analysis,Data curation, Methodology, Visualization, Writing – review & edit-ing, Writing – original draft, Supervision, Investigation, Resources,Software. HY: Formal analysis, Writing – original draft, Methodology,Software, Data curation, Writing – review & editing, Investigation.MA: Conceptualization, Data curation, Investigation, Methodology,Writing – review & editing, Writing – original draft, Visualization,Software, Formal analysis. AS: Resources, Conceptualization,Writing – original draft, Writing – review & editing, Supervision,Methodology. HB: Resources, Visualization, Project administration,Formal analysis, Writing – original draft, Investigation, Writing –review & editing, Data curation, Supervision, Methodology,Conceptualization. ### Page 13 Funding The author(s) declared that financial support was not received forthis work and/or its publication. Acknowledgments We are grateful to Namitha Thottiyil for assistance in participantrecruitment. Conflict of interest The author(s) declared that this work was conducted in theabsence of any commercial or financial relationships that could beconstrued as a potential conflict of interest. Generative AI statement The author(s) declared that Generative AI was not used in thecreation of this manuscript. Any alternative text (alt text) provided alongside figures in thisarticle has been generated by Frontiers with the support of artificialintelligence and reasonable efforts have been made to ensure accuracy,including review by the authors wherever possible. If you identify anyissues, please contact us. Publisher’s note All claims expressed in this article are solely those of the authorsand do not necessarily represent those of their affiliated organiza-tions, or those of the publisher, the editors and the reviewers. Anyproduct that may be evaluated in this article, or claim that may bemade by its manufacturer, is not guaranteed or endorsed by thepublisher. References Aghajani Mir, M. (2023). Brain fog: a narrative review of the Most common mysterious cogni-tive disorder in COVID-19. Mol. Neurobiol. 61, 9915–9926. doi: Aiyegbusi, O. L., Hughes, S. E., Turner, G., Rivera, S. C., McMullan, C., Chandan, J. S.,et al. (2021). Symptoms, complications and management of long COVID: a review. J. R.Soc. Med. 114, 428–442. doi: Aksoy, M., Hunter, S., Asghar, A. U. R., and Bhandari, S. (2024). A randomised con-trolled pilot trial protocol for patient led cognitive gamified training during haemodialy-sis. Sci. Rep. 14:30704. doi: Arbula, S., Pisanu, E., Bellavita, G., Menichelli, A., Lunardelli, A., Furlanis, G., et al.(2024). Insights into attention and memory difficulties in post-COVID syndrome usingstandardized neuropsychological tests and experimental cognitive tasks. Sci. Rep.14:4405. doi: World Medical Association (2013). World medical association declaration of Helsinki:ethical principles for medical research involving human subjects. JAMA 310, 2191–2194.doi: Atchison, C. J., Davies, B., Cooper, E., Lound, A., Whitaker, M., Hampshire, A., et al.(2023). Long-term health impacts of COVID-19 among 242,712 adults in England. Nat.Commun. 14:6588. doi: Baddeley, A. (2003). Working memory: looking back and looking forward. Nat. Rev.Neurosci. 4, 829–839. doi: Baddeley, A. D., Hitch, G. J., and Allen, R. J. (2019). From short-term store to multicom-ponent working memory: the role of the modal model. Mem. Cogn. 47, 575–588. doi: Baseler, H. A., Aksoy, M., Salawu, A., Green, A., and Asghar, A. U. R. (2022). The nega-tive impact of COVID-19 on working memory revealed using a rapid online quiz. PLoSOne 17:e0269353. doi: Bauer, L., Laksono, B. M., Vrij, F. M. S., Kushner, S. A., Harschnitz, O., and Riel, D.(2022). The neuroinvasiveness, neurotropism, and neurovirulence of SARS-CoV-2.Trends Neurosci. 45, 358–368. doi: Bertuccelli, M., Ciringione, L., Rubega, M., Bisiacchi, P., Masiero, S., and Del Felice, A.(2022). Cognitive impairment in people with previous COVID-19 infection: a scopingreview. Cortex 154, 212–230. doi: Blomberg, B., Mohn, K. G., Brokstad, K. A., Zhou, F., Linchausen, D. W., Hansen, B. A.,et al. (2021). Long COVID in a prospective cohort of home-isolated patients. Nat. Med.27, 1607–1613. doi: Bowe, B., Xie, Y., and Al-Aly, Z. (2022). Acute and postacute sequelae associ-ated with SARS-CoV-2 reinfection. Nat. Med. 28, 2398–2405. doi: Brockmole, J. R., and Logie, R. H. (2013). Age-related change in visual working memory:a study of 55,753 participants aged 8-75. Front. Psychol. 4:12. doi: Bungenberg, J., Humkamp, K., Hohenfeld, C., Rust, M. I., Ermis, U., Dreher, M., et al.(2022). Long COVID-19: objectifying most self-reported neurological symptoms. Ann.Clin. Transl. Neurol. 9, 141–154. doi: Chai, W. J., Abd Hamid, A. I., and Abdullah, J. M. (2018). Working memory from thepsychological and neurosciences perspectives: a review. Front. Psychol. 9:401. doi: Cheetham, N. J., Penfold, R., Giunchiglia, V., Bowyer, V., Sudre, C. H., Canas, L. S., et al.(2023). The effects of COVID-19 on cognitive performance in a community-basedcohort: a COVID symptom study biobank prospective cohort study. EClinicalMedicine62:102086. doi: Corbett, A., Williams, G., Creese, B., Hampshire, A., Hayman, V., Palmer, A., et al.(2023). Cognitive decline in older adults in the UK during and after the COVID-19pandemic: a longitudinal analysis of PROTECT study data. Lancet Healthy Longev. 4,e591–e599. doi: Cui, R., Gao, B., Ge, R., Li, M., Li, M., Lu, X., et al. (2024). The effects of COVID-19infection on working memory: a systematic review. Curr. Med. Res. Opin. 40, 217–227.doi: Davis, H. E., McCorkell, L., Vogel, J. M., and Topol, E. J. (2023). Long COVID: majorfindings, mechanisms and recommendations. Nat. Rev. Microbiol. 21, 133–146. doi: Diamond, A. (2013). Executive functions. Annu. Rev. Psychol. 64, 135–168. doi: Diexer, S., Klee, B., Gottschick, C., Xu, C., Broda, A., Purschke, O., et al. (2023).Association between virus variants, vaccination, previous infections, and post-COVID-19 risk. Int. J. Infect. Dis. 136, 14–21. doi: Elliott, P., Bodinier, B., Eales, O., Wang, H., Haw, D., Elliott, J., et al. (2022). Rapidincrease in omicron infections in England during December 2021: REACT-1 study.Science 375, 1406–1411. doi: Feltz-Cornelis, C., Turk, F., Sweetman, J., Khunti, K., Gabbay, M., Shepherd, J., et al.(2024). Prevalence of mental health conditions and brain fog in people with longCOVID: a systematic review and meta-analysis. Gen. Hosp. Psychiatry 88, 10–22. doi: Finamore, P., Arena, E., Lupoi, D., Savito, L., Nunzio, F., Furbatto, M., et al. (2024). LongCOVID syndrome: a. narrative review on burden of age and vaccination. J. Clin. Med.13:4756. doi: Finterer, J., and Scorza, F. A. (2022). A retrospective analysis of clinically confirmed longpost-COVID vaccination syndrome. J Clin Transl Res 8, 506–508. doi: Gao, Z., Xu, Y., Sun, C., Wang, X., Guo, Y., Qiu, S., et al. (2021). A systematic review ofasymptomatic infections with COVID-19. J. Microbiol. Immunol. Infect. 54, 12–16. doi: Gorenshtein, A., Liba, T., Leibovitch, L., Stern, S., and Stern, Y. (2024). Interventionmodalities for brain fog caused by long-COVID: systematic review of the literature.Neurol. Sci. 45, 2951–2968. doi: Graham, E. L., Clark, J. R., Orban, Z. S., Lim, P. H., Szymanski, A. L., Taylor, C., et al.(2021). Persistent neurologic symptoms and cognitive dysfunction in non-hospitalizedCovid-19 ‘long haulers’. Ann. Clin. Transl. Neurol. 8, 1073–1085. doi: Greenhalgh, T., Sivan, M., Perlowski, A., and Nikolich, J. Z. (2024). Long COVID: aclinical update. Lancet 404, 707–724. doi: Guo, P., Benito Ballesteros, A., Yeung, S. P., Liu, R., Saha, A., Curtis, L., et al. (2022).COVCOG 2: cognitive and memory deficits in long COVID: a second publication fromthe COVID and cognition study. Front. Aging Neurosci. 14:804937. doi: ### Page 14 Hampshire, A., Azor, A., Atchison, C., Trender, W., Hellyer, P. J., Giunchiglia, V., et al.(2024). Cognition and memory after Covid-19 in a large community sample. N. Engl. J.Med. 390, 806–818. doi: Hawke, L. D., Nguyen, A. T. P., Wang, W., Brown, E. E., Xu, D., Deuville, S., et al. (2024).Systematic review of interventions for mental health, cognition and psychological well-being in long COVID. BMJ Ment. Health 27, 1–11. doi: Jennings, G., Monaghan, A., Xue, F., Duggan, E., and Romero-Ortuno, R. (2022).Comprehensive clinical characterisation of brain fog in adults reporting long COVIDsymptoms. J. Clin. Med. 11:3440. doi: Kalak, G., Jarjou’i, A., Bohadana, A., Wild, P., Rokach, A., Amiad, N., et al. (2022).Prevalence and persistence of symptoms in adult COVID-19 survivors 3 and 18 monthsafter discharge from hospital or corona hotels. J. Clin. Med. 11:7413. doi: Katsuki, F. (2012). Unique and shared roles of the posterior parietal and dorsolateralprefrontal cortex in cognitive functions. Front. Integr. Neurosci. 6:17. doi: Kennelly, C. E., Nguyen, A. T. P., Sheikhan, N. Y., Strudwick, G., Ski, C. F.,Thompson, D. R., et al. (2023). The lived experience of long COVID: a qualitative studyof mental health, quality of life, and coping. PLoS One 18:e0292630. doi:[-](https://doi.org/10.1371/journal.pone.0292630) Kim, Y., Bae, S., Chang, H. H., and Kim, S. W. (2023). Long COVID prevalence andimpact on quality of life 2 years after acute COVID-19. Sci. Rep. 13:11207. doi: Lam, G. Y., Damant, R. W., Ferrara, G., Lim, R. K., Stickland, M. K., Ogando, N. S., et al.(2023). Characterizing long-COVID brain fog: a retrospective cohort study. J. Neurol.270, 4640–4646. doi: Lauria, A., Carfi, A., Benvenuto, F., Bramato, G., Ciciarello, F., Rocchi, S., et al. (2023).Neuropsychological measures of post-COVID-19 cognitive status. Front. Psychol.14:1136667. doi: Migeot, J., Calivar, M., Granchetti, H., Ibáñez, A., and Fittipaldi, S. (2022).Socioeconomic status impacts cognitive and socioemotional processes in healthy ageing.Sci. Rep. 12:6048. doi: Millet, C., Narvaneni, S., Roman, S., Horani, G., Chaudhry, S., Michael, P., et al. (2022).Symptoms persist in patients two years after COVID-19 infection: a prospective follow-up study. Clin. Microbiol. Infect. 28, 1505–1507. doi: Miskowiak, K. W., Johnsen, S., Sattler, S. M., Nielsen, S., Kunalan, K., Rungby, J., et al.(2021). Cognitive impairments four months after COVID-19 hospital discharge: pattern,severity and association with illness variables. Eur. Neuropsychopharmacol. 46, 39–48.doi: Miskowiak, K. W., Pedersen, J. K., Gunnarsson, D. V., Roikjer, T. K., Podlekareva, D.,Hansen, H., et al. (2023). Cognitive impairments among patients in a long-COVIDclinic: prevalence, pattern and relation to illness severity, work function and quality oflife. J. Affect. Disord. 324, 162–169. doi: Moller, M., Borg, K., Janson, C., Lerm, M., Normark, J., and Niward, K. (2023). Cognitivedysfunction in post-COVID-19 condition: mechanisms, management, and rehabilita-tion. J. Intern. Med. 294, 563–581. doi: Nalbandian, A., Sehgal, K., Gupta, A., Madhavan, M. V., McGroder, C., Stevens, J. S.,et al. (2021). Post-acute COVID-19 syndrome. Nat. Med. 27, 601–615. doi: Natarajan, A., Shetty, A., Delanerolle, G., Zeng, Y., Zhang, Y., Raymont, V., et al. (2023).A systematic review and meta-analysis of long COVID symptoms. Syst. Rev. 12:88. doi: NHS. (2021). Post-COVID Syndrome Long COVID National Health Service England.Available online at: (Accessed May 13, 2026). NICE. (2020). COVID-19 Rapid Guideline: Managing the Long-Term Effects of COVID-19National Institute for Health and Care Excellence. Available online at: (Accessed May 13, 2026). Nordvig, A. S., Rajan, M., Lau, J. D., Kingery, J. R., Mahmud, M., Chiang, G. C., et al.(2023). Brain fog in long COVID limits function and health status, independently ofhospital severity and preexisting conditions. Front. Neurol. 14:1150096. doi: O’Mahoney, L. L., Routen, A., Gillies, C., Ekezie, W., Welford, A., Zhang, A., et al. (2023).The prevalence and long-term health effects of long covid among hospitalised and non-hospitalised populations: a systematic review and meta-analysis. EClinicalMedicine55:101762. doi: Office for National Statistics. (2023). Coronavirus (COVID-19) Latest Insights: Infections.Newport: Office for National Statistics. Available online at:. (Accessed May 13, 2026). Park, D. C., Lautenschlager, G., Hedden, T., Davidson, N. S., Smith, A. D., andSmith, P. K. (2002). Models of visuospatial and verbal memory across the adult life span.Psychol. Aging 17, 299–320. doi: Plantone, D., Stufano, A., Righi, D., Locci, S., Iavicoli, I., Lovreglio, P., et al. (2024).Neurofilament light chain and glial fibrillary acid protein levels are elevated in post-mildCOVID-19 or asymptomatic SARS-CoV-2 cases. Sci. Rep. 14:6429. doi: Premraj, L., Kannapadi, N. V., Briggs, J., Seal, S. M., Battaglini, D., Fanning, J., et al.(2022). Mid and long-term neurological and neuropsychiatric manifestations of post-COVID-19 syndrome: a meta-analysis. J. Neurol. Sci. 434:120162. doi: Repovs, G., and Baddeley, A. (2006). The multi-component model of working memory:explorations in experimental cognitive psychology. Neuroscience 139, 5–21. doi: Salawu, A., Green, A., Crooks, M. G., Brixey, N., Ross, D. H., and Sivan, M. (2020). Aproposal for multidisciplinary tele-rehabilitation in the assessment and rehabilitation ofCOVID-19 survivors. Int. J. Environ. Res. Public Health 17. doi: Schroeder, R. W., Twumasi-Ankrah, P., Baade, L. E., and Marshall, P. S. (2012). Reliabledigit span: a systematic review and cross-validation study. Assessment 19, 21–30. doi: Stavem, K., Einvik, G., Tholin, B., Ghanima, W., Hessen, E., and Lundqvist, C. (2022).Cognitive function in non-hospitalized patients 8-13 months after acute COVID-19infection: a cohort study in Norway. PLoS One 17:0273352. doi: Tanabe, A., and Osaka, N. (2009). Picture span test: measuring visual working memorycapacity involved in remembering and comprehension. Behav. Res. Methods 41,309–317. doi: Tedjasukmana, R., Budikayanti, A., Islamiyah, W. R., Witjaksono, A., and Hakim, M.(2022). Sleep disturbance in post COVID-19 conditions: prevalence and quality of life.Front. Neurol. 13:1095606. doi: Thronicke, A., Hinse, M., Weinert, S., Jakubowski, A., Grieb, G., and Matthes, H. (2022).Factors associated with self-reported post/long-COVID-A real-world data study. Int. J.Environ. Res. Public Health 19:16124. doi: Turner, S., Khan, M. A., Putrino, D., Woodcock, A., Kell, D. B., and Pretorius, E. (2023).Long COVID: pathophysiological factors and abnormalities of coagulation. TrendsEndocrinol. Metab. 34, 321–344. doi: WHO. (2022). Post COVID-19 Condition Long COVID World Health Organization.Available online at: (Accessed May 13, 2026). Zawilska, J. B., and Kuczynska, K. (2022). Psychiatric and neurological complications oflong COVID. J. Psychiatr. Res. 156, 349–360. doi: Zhao, S., Toniolo, S., Hampshire, A., and Husain, M. (2023). Effects of COVID-19 oncognition and brain health. Trends Cogn. Sci. 27, 1053–1067. doi: Ziauddeen, N., Gurdasani, D., O’Hara, M. E., Hastie, C., Roderick, P., Yao, G., et al.(2022). Characteristics and impact of long covid: findings from an online survey. PLoSOne 17:e0264331. doi: