Executive functions and processing speed in covert cerebral small vessel disease

Abstract

Background and purpose

Executive dysfunction and slowed processing speed are central cognitive impairments in cerebral small vessel disease (cSVD). It is unclear whether the subcomponents of executive functions become equally affected and whether computerized tests are more sensitive in detecting early cognitive changes over traditional tests. The associations of specific executive abilities (cognitive flexibility, inhibitory control, working memory) and processing speed with white matter hyperintensities (WMHs) and Instrumental Activities of Daily Living (IADL) were examined.

Methods

In the Helsinki Small Vessel Disease Study, 152 older individuals without stroke or dementia were assessed with brain magnetic resonance imaging and comprehensive neuropsychological evaluation. WMH volumes were obtained with automated segmentation. Executive functions and processing speed measures included established paper‐and‐pencil tests and the computer‐based Flexible Attention Test (FAT), Simon task and Sustained Attention to Response Task.

Results

White matter hyperintensity volume and IADL were associated with multiple cognitive measures across subdomains independently of demographic factors. The highest effect sizes were observed for FAT numbers and number‐letter tasks (tablet modifications from the Trail Making Test), FAT visuospatial span, Simon task and semantic verbal fluency. Some of the widely used tests such as Stroop inhibition, phonemic fluency and digit span were not significantly associated with either WMHs or IADL.

Conclusion

Processing speed and executive function subcomponents are broadly related to functional abilities and WMH severity in covert cSVD, but the strength of associations within subdomains is heavily dependent on the assessment method. Digital tests providing precise measures of reaction times and response accuracy seem to outperform many of the conventional paper‐and‐pencil tests.

INTRODUCTION

Cerebrovascular disease and cerebral small vessel disease (cSVD), in particular, are major causes of cognitive and functional impairment in older people. cSVD is characterized by diverse brain changes, such as white matter hyperintensities (WMHs) and lacunes of presumed vascular origin, microbleeds, perivascular spaces and brain atrophy [1, 2], which are often incidentally discovered on neuroimaging without clear onset of clinical symptoms. ‘Covert’ cSVD without apparent neurological manifestations, such as stroke or dementia, has been identified as a critical target for management to preserve functional abilities and well‐being in older age [3, 4]. With the rapid evolution of knowledge on imaging biomarkers, the diagnosis of cSVD is primarily based on magnetic resonance imaging (MRI) findings. The advancement of effective cognitive markers has received less attention, even though cognitive decline is one of the main clinical outcomes in cSVD.

Cerebral SVD is related to impairment in several cognitive domains including processing speed, attention and executive functions, as well as memory, reasoning, visuospatial abilities and language [5]. With progressing brain changes cognitive functioning becomes globally affected, but it is unclear whether this decline is initially driven by domain‐specific deficits. The imaging findings of cSVD have been most strongly associated with executive dysfunction and slowed processing speed [6, 7, 8, 9, 10, 11, 12], and these impairments may lead to decline in other cognitive domains such as episodic memory and visuospatial functions [13, 14], although the specificity of these impairments is debated [5, 12, 15].

Executive functions, as an umbrella concept, consist of a variety of subcomponents. Yet, in cSVD research, these abilities are often evaluated with single tests or composite scores averaging multiple test results together. There are several conceptual taxonomies for executive functions, one of the best known being the data‐driven model of Miyake et al., which classifies executive functions into components of set‐shifting (cognitive flexibility), inhibition and updating (working memory) [16]. These components are separable although overlapping abilities closely linked with speed of information processing and complex attention.

Currently, there is no consensus on assessment methods for the identification of cognitive impairment in cSVD, and no single test has the ability to differentiate cSVD from other neurocognitive disorders [12]. According to a recent meta‐analysis, the most commonly used cognitive tests to assess attention and executive functions in cSVD research have been the Stroop, Trail Making and verbal fluency tests, which as timed measures have shown the best potential to differentiate cSVD from healthy controls [17]. These traditional tests, dating back to the mid‐20th century, are established in clinical neuropsychological practice in many parts of the world. However, digitalized technologies have challenged the classic paper‐and‐pencil methods by improving stimulus presentation, measurement accuracy and inter‐tester reliability [18, 19].

Based on the current literature, it is unclear whether specific executive abilities are differentially affected by cSVD brain changes, and whether computerized tests are more sensitive in detecting deficits than paper‐and‐pencil tests. This study investigated the associations of WMH volume with processing speed, cognitive flexibility, inhibitory control and working memory by using an extensive set of conventional and computer‐based neuropsychological tests. To verify the clinical significance of the findings, the relationship of processing speed and executive subcomponents with Instrumental Activities of Daily Living (IADL) was concurrently examined.

MATERIALS AND METHODS

Participants and study protocol

Baseline data of 152 participants from the Helsinki Small Vessel Disease Study, a prospective cohort study investigating the evolution of cognitive, clinical and imaging features in early‐stage cSVD, were analysed. All participants gave their written informed consent, and the study was approved by the Ethics Committee of Helsinki University Hospital, Finland. The participants were recruited from the Helsinki University Hospital imaging registry based on any degree of WMHs as detected in routine brain imaging. All participants were invited to another study‐specific brain MRI and clinical assessments including medical, functional and neuropsychological examinations with standard protocol [20, 21]. The average time between the study MRI and neuropsychological assessment was 29 days (min 4, max 103).

Inclusion criteria to the study enrolment were age 65–75 years; residence within the Helsinki and Uusimaa Hospital District, Finland; occurrence of no more than minor, temporary and local neurological symptoms (manifested 3–12 months before the enrolment) or no neurological symptoms; no more than slight disability in basic daily activities defined by the modified Rankin Scale; and sufficient hearing, vision and Finnish language skills to complete the assessments. Exclusion criteria were significant neurological disease (including symptomatic stroke and dementia according to the Diagnostic and Statistical Manual of Mental Disorders, fifth edition); severe psychiatric disorder; current substance abuse; severe other medical condition preventing participation; traumatic brain injury that has required hospitalization; severe intellectual disability; and inability or refusal to undergo brain MRI. Additional exclusion criteria based on MRI findings were cortical infarct; subcortical infarct larger than 15 mm (20 mm on diffusion‐weighted images); haemorrhage larger than microbleed (>10 mm); brain tumour; and contusion, traumatic haemorrhage or diffuse axonal injury.

Magnetic resonance imaging

Brain MRI was performed using a 3 T scanner with 32‐channel head coil (Siemens Magnetom Skyra/Verio). The following sequences were included: a fast T1 gradient echo localizer in three orthogonal directions, sagittal 3D fluid‐attenuated inversion recovery (FLAIR) sampling perfection with application optimized contrast (SPACE), sagittal 3D T2 SPACE, 3D gradient echo susceptibility weighted imaging sequence, and sagittal 3D T1 magnetization prepared rapid gradient echo imaging (MPRAGE) [21]. WMHs were selected as the main MRI surrogate of cSVD severity based on their strong associations with cognitive and functional symptom progression [22]. WMHs were determined according to the Standards for Reporting Vascular Changes on Neuroimaging as hyperintense areas in the white matter without cavitation on FLAIR sequences [1, 2]. At first, an experienced neuroradiologist assessed WMH visually and classified the degree according to the modified Fazekas scale [23] for descriptive purposes. For the main analyses, WMH volume was determined with the fully automated cNeuro image quantification tool (www.cneuro.com/cmri/Combinostics, Finland) [24, 25]. WMHs were segmented on FLAIR images according to a multi‐stage method based on the expectation–maximization algorithm [26]. The total WMH volume was normalized for intracranial volume [27].

Neuropsychological assessment

A comprehensive neuropsychological assessment battery was administered with tests covering processing speed, executive functions, memory and learning, visuospatial abilities and verbal functions [21]. The assessment was performed by trained psychologists in a quiet testing room (duration approximately 2 h). The present study focused on the in‐depth analysis of processing speed and executive functions subdomains including cognitive flexibility, inhibitory control and working memory. The evaluation of these functions consisted of established paper‐and‐pencil tests alongside computer‐based cognitive tasks, namely the Flexible Attention Test (FAT) [28], Simon task [29, 30] and Sustained Attention to Response Task (SART) [30, 31], described below and illustrated in Figures 1, 2, 3.

FIGURE 1.

The Flexible Attention Test (FAT) consists of eight subtasks with progressing difficulty evaluating visuomotor speed, attention, set‐shifting and cognitive flexibility, and visuospatial working memory. The test is an extended modification from the classic Trail Making Test and Corsi Block‐Tapping Test and is carried out on a touch‐screen tablet computer, taking altogether 10–15 min to complete. Each subtask consists of 24 stimuli randomly located on the screen.

FIGURE 2.

The Simon task is a computer‐based measure of response inhibition and interference conflict resolution. A blue or red square is presented either on the left or the right side of the screen. The subjects are asked to respond by pressing the right button to a red square and the left button to a blue square. In congruent trials the square is on the same side as the expected response and in incongruent trials it is on the opposite side.

FIGURE 3.

The Sustained Attention to Response Task (SART) is a computer‐based go/no‐go task that requires subjects to withhold response to an infrequent target presented amongst frequent non‐targets. Digits ranging from 1 to 9 are presented at the centre of the screen in a random order at a brisk pace. The task is to press the response button as quickly as possible for all other digits except ‘3’.

Processing speed was evaluated with the Wechsler Adult Intelligence Scale IV (WAIS‐IV) coding subtest [32], the Stroop test colour‐congruent condition [33] and the FAT reaction time and numbers subtasks. Cognitive flexibility (initiation and set‐shifting) was assessed with the verbal fluency test (semantic and phonemic) [34], the Brixton Spatial Anticipation Test [35] and the FAT set‐shifting subtasks. Inhibitory control was evaluated with the Stroop test colour‐incongruent condition [33], Hayling Sentence Completion Test [35], Simon task and SART. Working memory was assessed with the Wechsler Memory Scale III (WMS‐III) digit span and number‐letter sequencing subtasks [36] and the FAT visuospatial span subtask.

FAT is a novel extended modification based on the classic Trail Making Test [37] and Corsi Block‐Tapping Test [38], developed at the Finnish Institute of Occupational Health [28]. It consists of eight subtasks with progressing difficulty measuring visuomotor speed, attention, cognitive flexibility and visuospatial working memory, taking approximately 10–15 min in total (Figure 1). Each subtask has the same number of visual stimuli presented at semi‐randomized locations (keeping the same path length) to reduce practice effect in repeated assessments. FAT was administered on a 12.3‐inch touch‐screen tablet computer with Windows 10 Pro 64‐bit operating system as detailed in Tikkanen et al. [28]. The Simon task is a measure of conflict resolution and response inhibition and requires the participant to respond to a non‐spatial feature (colour) of lateralized stimuli by pressing either of two horizontally placed response buttons in congruent and incongruent conditions (Figure 2) [29, 30, 39]. SART is a go/no‐go task that requires participants to withhold response to an infrequent target (1 in 9) presented amongst frequent non‐targets and measures sustained attention and inhibition to irrelevant stimuli (Figure 3 ) [30, 31]. Both the Simon task and SART were administered on the tablet computer using Neurobehavioral Systems Presentation software and an external response pad (Cedrus RB‐540). Each computer test condition was preceded by a short practice phase. Details of the computer test variables and their correlations with paper‐and‐pencil tests are given in Table S1.

Evaluation of IADL

Functional abilities were evaluated by the participant's close informant (spouse, other family member or close friend) with the Amsterdam Instrumental Activities of Daily Living (A‐IADL) questionnaire [40]. The short A‐IADL questionnaire includes 30 questions related to complex everyday activities such as household duties, operating appliances, handling finances and using public transport, and has been associated with WMHs in cSVD [21]. The responses are given on a five‐point scale focusing on activities within the last 4 weeks. The total score is determined using item response theory, with higher scores reflecting less impairment.

Statistical analysis

The results were analysed with multivariable linear regression models adjusted for participant's age, gender and years of education. WMH volume or IADL score were used as independent variables and cognitive test scores as dependent variables. Individual test scores were used to differentiate between specific aspects of executive function subcomponents and to identify the measures with the strongest associations with WMHs and IADL. In addition, combined scores were calculated for the FAT subtasks of visuomotor speed (reaction time and numbers), set‐shifting (the four set‐shifting tasks) and working memory (visuospatial span forward and backward) by averaging the standardized individual scores.

Because of skewed distributions, log transformation was applied to WMH volume and square root or log transformations to cognitive scores, where appropriate. There were a few missing values in the cognitive variables due to the participant's inability to complete the test, sporadic technical malfunction in computerized tests or other reasons. Missing data were not replaced, and therefore the number of observations in each analysis slightly varied (available case analysis). WAIS‐IV coding, Stroop, verbal fluency, Hayling test and WMS‐III number‐letter sequencing, and digit span subtests had no missing values. The highest number of missing values was found in the FAT visuospatial span backward task (n = 18). Participants with missing data in this test did not differ from those with available data in demographic variables or the Mini‐Mental State Examination (MMSE) score (p > 0.05). Moreover, informant evaluations with the A‐IADL questionnaire were available for 132 participants. Participants with missing A‐IADL data (n = 20) were more often men (p = 0.026), but they did not differ from the rest of the sample in age, education, WMH volume or MMSE score (p > 0.05). Statistical significance with the conventional level p < 0.05 was reported and the effect sizes between outcomes were compared with Cohen's f [2].

RESULTS

Characteristics of participants

Demographic characteristics, vascular risk factors and WMH findings of the study participants are presented in Table 1. The mean MMSE score was 27.4 (SD 2.3), and 133 (88%) participants scored above the cut‐off for impairment (<25/30). Descriptive data of the cognitive scores in the total sample and in three WMH severity groups are given in Table S3.

TABLE 1.

Characteristics of the participants (n = 152).

Demographics
Age, years, mean (SD)	70.6 (2.9)
Men/women, n	57/95
Education, years, mean (SD)	13.0 (4.5)
Referral reasons for initial brain imaging
Transient ischaemic attack, n (%)	40 (26)
Dizziness, n (%)	28 (18)
Headache or migraine, n (%)	15 (10)
Subjective cognitive complaints, n (%)	6 (4)
Visual symptoms, n (%)	17 (11)
Fall, n (%)	11 (7)
Syncope, n (%)	4 (3)
Other reasons (e.g., changes in gait and balance, sensory deficits), n (%)	31 (20)
Vascular risk factors
Hypertension, n (%)	105 (70)
Diabetes, n (%)	33 (22)
Hypercholesterolaemia, n (%)	118 (78)
Atrial fibrillation, n (%)
17 (11)
Current smoking, n (%)	10 (7)
White matter hyperintensities
Fazekas score, no/mild/moderate/severe, n	14/76/42/20
Total volume, mL a , mean, SD	10.5 (13.8)

^a Normalized for intracranial volume.

Associations between cognitive performance, WMHs and IADL

The main results are summarized in Table 2. Linear regression analyses adjusted for age, gender and education showed significant associations of both WMH volume and A‐IADL score with cognitive performances within all evaluated subdomains. However, the strength of associations varied between different tests from small to nearly medium according to Cohen's criterion [41]. Within processing speed tests, the highest effect sizes were observed in FAT numbers and FAT visuomotor speed combined score, followed by WAIS‐IV coding and the Stroop colour‐congruent condition. Within measures of cognitive flexibility, the strongest effects were seen in FAT numbers and letters, FAT numbers and months forward, FAT set‐shifting combined score and semantic verbal fluency. For the inhibitory control tests the highest effect size was found in the Simon task, whereas for working memory the strongest effects were in FAT visuospatial span forward and working memory combined score. The associations between A‐IADL score and cognitive scores were similar to those of WMH volume, but generally somewhat weaker. Brixton Spatial Anticipation Test, phonemic fluency, Stroop test colour‐incongruent condition, WMS‐III digit span, WMS‐III number‐letter sequencing, FAT numbers and months backward, and FAT reaction time had no significant associations with either WMH volume or the A‐IADL score.

TABLE 2.

Associations of WMHs and IADL with processing speed and executive function subdomains.

Dependent variable	WMH volume		IADL score
	Stand. β (p)	Effect size f 2	Stand. β (p)	Effect size f 2
Processing speed
FAT visuomotor speed score	0.25 (0.002)	0.07	−0.31 (<0.001)	0.11
FAT reaction time, time	0.14 (0.077)	0.02	−0.07 (0.392)	0.01
FAT numbers, time	0.28 (<0.001)	0.09	−0.33 (<0.001)	0.13
WAIS‐IV coding	−0.20 (0.009)	0.05	0.17 (0.040)	0.03
Stroop test, colour‐congruent, time	0.21 (0.008)	0.05	−0.16 (0.073)	0.03
Cognitive flexibility
FAT set‐shifting score	0.32 (<0.001)	0.13	−0.20 (0.015)	0.05
FAT numbers and letters, time/correct	0.34 (<0.001)	0.14	−0.19 (0.025)	0.04
FAT numbers and shapes, time/correct	0.26 (0.002)	0.07	−0.14 (0.104)	0.02
FAT numbers and months forward, time/correct	0.23 (0.007)	0.06	−0.27 (0.002)	0.09
FAT numbers and months backward, time/correct	0.16 (0.057)	0.03	−0.07 (0.409)	0.01
Verbal fluency, semantic (animals)	−0.28 (<0.001)	0.09	0.21 (0.013)	0.05
Verbal fluency, phonemic (letters K and S)	−0.09 (0.259)	0.01	0.02 (0.823)	0.00
Brixton Spatial Anticipation Test, errors	0.14 (0.087)	0.02	−0.10 (0.252)	0.01
Inhibitory control
Simon task, percentage correct (incongruent)	−0.23 (0.006)	0.06	0.26 (0.003)	0.07
Sustained Attention to Response Task, error rate	0.17 (0.044)	0.03	−0.10 (0.276)	0.01
Hayling Sentence Completion Test, error score	0.16 (0.047)	0.03	−0.15 (0.071)	0.03
Stroop test, colour‐incongruent, time/correct	0.14 (0.074)	0.02	−0.08 (0.368)	0.01
Working memory
FAT working memory score	−0.28 (0.001)	0.08	0.18 (0.048)	0.03
FAT visuospatial span forward	−0.23 (0.009)	0.05	0.20 (0.032)	0.04
FAT visuospatial span backward	−0.21 (0.018)	0.04	0.11 (0.231)	0.01
WMS‐III digit span	−0.05 (0.516)	0.00	0.04 (0.657)	0.00
WMS‐III number‐letter sequencing	−0.13 (0.094)	0.02	0.09 (0.285)	0.01

Note: Multivariable linear regression analyses adjusted for age, gender and years of education. Standardized β coefficients (p values) and Cohen's f ² for effect size.

Abbreviations: FAT, Flexible Attention Test; IADL, Instrumental Activities of Daily Living; WAIS‐IV, Wechsler Adult Intelligence Scale IV; WMHs, white matter hyperintensities; WMS‐III, Wechsler Memory Scale III.

DISCUSSION

Slowed processing speed and executive dysfunction are considered hallmark cognitive symptoms in cSVD, but the specific nature of these deficits has been poorly characterized. Thus far in cSVD research, executive impairments have often been bundled into a single entity, although executive functions, as a multidimensional construct, comprise several subcomponents of complex goal‐directed behaviour [16, 34]. No single test has the ability to cover the entire spectrum of executive abilities, and therefore their reliable assessment requires a comprehensive approach. In the present study, an extensive battery of neuropsychological paper‐and‐pencil tests and computerized measures was leveraged to analyse the associations of WMH burden with processing speed and executive subdomains in patients with covert cSVD (no clinical stroke or dementia). The aim was to identify the most promising tools for symptom identification and validate the results against everyday functional abilities.

Overall, the results showed that more severe WMH burden was linked to a broad range of difficulties in executive abilities and processing speed, but there were considerable differences between the individual measures in their ability to detect these impairments.

Specifically, within the processing speed domain, the strongest associations of WMHs were observed with the FAT numbers subtask, which is a tablet computer modification of the Trail Making A measuring visuomotor speed and attention [34, 42]. The paper‐and‐pencil tests, WAIS‐IV coding and Stroop colour‐congruent condition, evaluating processing speed in visuomotor and verbal modalities, respectively, also reached significance, but with slightly lower effect sizes. Notably, WMHs were not associated with FAT reaction time subtask, which as a simple motor speed task places fewer demands on attentional regulation compared to numbers subtask.

Within the cognitive flexibility subdomain, the FAT set‐shifting score, particularly the numbers and letters subtask, was found to be most strongly associated with WMHs. These set‐shifting subtasks are a series of tablet modifications from the Trail Making B with extended stimulus variability. In addition to numbers and letters, FAT incorporates months (forward or backward) and shapes (circles and squares), yet the basic idea in each task remains the same, that is, the subject needs to simultaneously keep two streams of thought and switch action between stimulus types [42]. In addition, WMHs were significantly associated with semantic fluency reflecting flexible verbal productivity. Previous studies have also shown cSVD brain changes to be associated with deterioration of semantic fluency [43]. Unexpectedly, no link between WMHs and phonemic fluency was found, even though it is considered primarily a measure of executive processing, whereas semantic fluency is sometimes grouped under the language domain. Both types of verbal fluency tasks, however, essentially relate to strategic word retrieval, self‐monitoring and set‐shifting, and strongly correlate with each other [44]. Lastly, within the cognitive flexibility domain, WMHs were not associated with the Brixton Spatial Anticipation Test assessing the ability to detect and follow a rule, and switch to a new rule [35]. Brixton involves similar cognitive processes to the Wisconsin Card Sorting Test but is somewhat shorter and easier, which may make it less sensitive to mild deficits.

As measures of inhibitory control, the paper‐and‐pencil versions of the Stroop test and Hayling Sentence Completion Test were used, as well as the computer‐based Simon task and SART. WMHs had the strongest relation with the percentage of correct responses in the incongruent condition of the Simon task, in which the subject is asked to selectively respond to specific qualities of the stimuli and inhibit responses to others [39]. WMHs were significantly, although weakly, also associated with the SART go/no‐go task and Hayling error score (inhibition lapses). However, contrary to our expectations, the Stroop colour‐incongruent condition did not reach significance. The Stroop test has been one of the most popular executive tests in cSVD and ageing research [17, 45] and regarded as a sensitive measure of cognitive decline, although conflicting results have been shown also before [43]. Several versions and scoring systems have been proposed for the Stroop test [34, 46]. Here, the time to complete the test divided by the number of correct responses was used to indicate both speed of performance and inhibition errors in the incongruent condition. Both the Simon and Stroop tests are timed assessments; however, the Simon task runs at a fast, forced pace, whereas in the Stroop test the subjects proceed at their own pace.

In the working memory subdomain, the FAT visuospatial span scores were the sole measures linked to WMHs with relatively robust effect sizes. Conversely, the digit span and number‐letter sequencing subtests from the Wechsler scales yielded no significant associations. Thus, WMH volume was only associated with the visuospatial component of working memory, rather than its verbal counterpart. This distinction might be attributable to the differing levels of task difficulty, as well as the specific modality involved. The FAT working memory task has been adapted for a touch‐screen tablet from the Corsi Block‐Tapping Test and features a larger number of ‘blocks’ (24 circles) that appear in semi‐randomly varying locations (Figure 1). This format is likely to place greater demands on working memory and reduces the possibility of relying on long‐term memory strategies. Similar to the Corsi approach, the sequences to be recalled incrementally increase in length, starting with two items, and the task is terminated after two consecutive failures in the same span length [38].

To shed light on the ecological validity of the above‐described findings, the associations of processing speed and executive performances with everyday functional abilities as evaluated with an informant‐based Amsterdam IADL questionnaire were further examined. This method has good psychometric qualities and measurement accuracy along the spectrum functional abilities from normal ageing to dementia [40]. Convincingly, a very similar pattern of significant associations between cognitive measures and IADL was found to those between cognition and WMH volume. Although the relationships were slightly weaker compared to WMHs, IADL was significantly associated with a wide range of cognitive subdomains. Based on effect sizes, the most sensitive cognitive measures were the FAT processing speed and set‐shifting scores, Simon task and semantic fluency. Again, many of the conventional tests such as the Stroop test, phonemic fluency, Brixton and Hayling tests, digit span and number‐letter sequencing were non‐significant.

Taken together, the results suggest that computer‐based tests have great potential in detecting mild deficits in processing speed and executive function subcomponents in cSVD. This is not surprising, given the advantages of digital methods in terms of time accuracy and manifold stimulus presentation possibilities. Cognitive tools sensitive to the decrease of processing speed have been suggested to be most promising in defining a clinical profile of cSVD [12]. The benefits of computer‐based tests also include invariable administration, adaptivity according to individual performance level and quick automated scoring. On the other hand, computer methods may be limited by insufficient information on psychometric properties, lack of normative data and technical obstacles [18, 47, 48]. It is also important to note that computer and paper‐and‐pencil tests do not measure the same aspects of cognitive function, even within the same subdomain. Due to these challenges, the clinical community has been somewhat conservative in adopting computerized methods into clinical practices [47, 49]. Obviously, computer‐based tests are not automatically superior to pencil‐and‐paper tasks and should meet the same standards for the development and use as any other neuropsychological test [18]. More research is needed on the reliability and validity as well as norms and suitability of digital tests for specific target groups.

Another obvious challenge for the evaluation of executive functions comes from the complex multidimensional nature of the construct and the overlap of its subcomponents with other cognitive domains. Neuropsychological measures typically tap multiple aspects of executive function together with non‐executive abilities [50]. In the end, the most reliable approach in clinical neuropsychological assessment is probably achieved by a combination of digital and pen‐and‐paper methods.

The strengths of the present study include an extensive battery of both conventional and innovative neuropsychological tests, allowing for a comprehensive subcomponent‐level characterization of processing speed and executive deficits. It was possible to examine the associations of these functions with volumetric WMH data, ranging from minimal to severe, in a covert (subclinical) stage of cSVD. Furthermore, the clinical significance of the findings was validated in relation to everyday functional abilities, as assessed by the participants' close informants. The absence of a representative control group of neurologically completely healthy older individuals prevented us from presenting normative data or specific cut‐points for distinguishing normal versus abnormal performances. Although the sample consisted of relatively young elderly people, in whom mixed aetiologies are less common, the possibility of concomitant neurodegenerative pathologies cannot be ruled out. The study is also limited by its sample size and cross‐sectional design, thus representing exploratory research on processing speed and executive dysfunction in cSVD.

In conclusion, cognitive markers of processing speed and executive function subdomains were differentially related to WMHs and IADL in covert cSVD. WMH volume was consistently associated with a wide range of subcomponents, but the strength of associations within subdomains was highly dependent on the evaluation method. Computerized tests providing precise measures of reaction times and response accuracy seemed to outperform many of the conventional paper‐and‐pencil tests. These findings underscore the importance of choosing sensitive outcome measures in studies on cognitive impairment in cSVD. Developing effective cognitive markers is essential for early detection and timely intervention before the onset of overt dementia.

AUTHOR CONTRIBUTIONS

Hanna Jokinen: Conceptualization; funding acquisition; writing – original draft; methodology; visualization; formal analysis; data curation; supervision; investigation; project administration; validation; resources. Hanna M. Laakso: Writing – review and editing; investigation; conceptualization; methodology; formal analysis. Anne Arola: Investigation; writing – review and editing; methodology; formal analysis. Teemu I. Paajanen: Software; visualization; writing – review and editing; methodology. Jussi Virkkala: Software; writing – review and editing; methodology. Teppo Särkämö: Software; writing – review and editing; methodology. Tommi Makkonen: Software; writing – review and editing; visualization; methodology. Iiris Kyläheiko: Writing – review and editing; formal analysis. Heidi Heinonen: Investigation; writing – review and editing. Johanna Pitkänen: Writing – review and editing; investigation; formal analysis; data curation; methodology. Antti Korvenoja: Methodology; investigation; supervision; conceptualization; writing – review and editing. Susanna Melkas: Conceptualization; investigation; funding acquisition; writing – review and editing; methodology; project administration; supervision; resources.

FUNDING INFORMATION

The study was funded by the Research Council of Finland, Helsinki and Uusimaa Hospital District, and University of Helsinki.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Data S1.

Jokinen H, Laakso HM, Arola A, et al. Executive functions and processing speed in covert cerebral small vessel disease. Eur J Neurol. 2025;32:e16533. doi: 10.1111/ene.16533

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.