Participants

The study included adults (over 18 years of age), of both sexes, divided into three groups: a) 45 neurologically healthy individuals, b) 11 with post-stroke RHD, and c) 38 with post-stroke LHD. The final sample consisted of 94 adults, 52 of them female. The healthy, RHD, and LHD groups did not show statistically significant differences in sex, age, and years of education. The post-stroke patients exhibited a heterogeneous profile based on the types and locations of their lesions. Table 1 provides descriptive data for each group.

All participants were Brazilian, monolingual, and had at least four years of formal education. For the healthy group, individuals without any history of stroke or other neurological conditions were included. Post-stroke patients could have up to two episodes of lesions in the same cerebral hemisphere (right or left). They should not have neurological damage resulting from other diseases, such as tumors or traumatic brain injuries, confirmed by neuroimaging exams and neurological assessments.

Exclusion criteria for all groups (healthy and post-stroke) included a current or previous history of substance abuse (alcohol and illicit drugs) and self-reported psychiatric or sensory disorders (uncorrected hearing and/or visual impairments). Adults with moderate to severe depressive symptoms, as indicated by their responses on the depression scales selected according to the participant’s age (described in the instruments section), were also excluded. Additionally, for the neurologically healthy group, evidence of cognitive decline assessed through the Mini-Mental State Examination (^{Folstein et al., 1975}) was an exclusion criterion, following the cutoff points proposed by ^{Kochhann et al. (2010)}, which vary according to education level.

Data collection procedures and Instruments

The study participants were part of a project that obtained approval from the Research Ethics Committee of a general hospital in Porto Alegre, under number 100149, and of a public university in Rio Grande do Sul (authorization number 2009028). All participants signed the consent form. The stroke patients were contacted by telephone after being referred by doctors and residents at the general hospital or after an analysis of their medical records. The neurologically healthy participants were from the general community. The following instruments were administered:

a) Health and sociodemographic conditions questionnaire: produced by the research group, consisting of questions about age, years of education, gender, substance use history, and history of psychiatric and neurological diagnoses (sensory and others). The questionnaire was used for sample characterization and to relate data such as age and education to neuropsychological performance.

b) Brief Neuropsycholinguistic Assessment Instrument - NEUPSILIN-L (^{Fontoura et al., 2011}): this instrument consists of a battery of 33 neuropsychological tasks, assessed with options for oral responses (OR), motor responses (MR), or both. The instrument assesses temporal-spatial orientation (OR - 8 points and MR - 8 points), attention (reverse counting - 20 points OR; digit sequence repetition - 7 points OR and 7 points MR), visual perception (verification of line equalities and differences - 6 points; visual hemineglect - 1 point; face perception - 3 points; face recognition - 2 points), working memory (backwards digit ordering - 10 points MR; auditory word span in sentences - 28 points MR), episodic-semantic verbal memory (immediate, delayed, and recognition - 40 points OR), long-term semantic memory (5 points OR and 5 points MR), short-term visual memory (3 points), arithmetic skills (8 points), oral language (naming - 4 points; repetition - 10 points; comprehension - 3 points; inference processing - 3 points OR and 3 points MR), written language (reading - 12 points; comprehension - 3 points; spontaneous writing - 2 points; copied writing - 2 points; dictated writing - 12 points), praxis (ideomotor - 3 points; constructive - 16 points; reflective - 3 points), and executive functions (problem-solving - 2 points OR and 2 points MR; orthographic and semantic verbal fluency - score dependent on the number of words evoked). Tasks that assess neuropsychological functions visually do not have OR options. Tasks that depend on speech are answered in both OR and multiple-choice (MR) forms, always in this order. Tasks of verbal memory and language do not have MR options since this form of response would alter the target function being assessed. The NEUPSILIN-L scoring is done separately based on the number of correctly verbalized and pointed items, with scores extracted for both tasks and functions (task groups).

c) Boston Aphasia Diagnostic Test - Short version (^{Goodglass, Kaplan, & Barresi, 2001}; ^{Radanovic, Mansur, Azambuja, Porto, & Scaff, 2004)}: applied exclusively to the post-stroke adult group by a speech therapist experienced in clinical diagnosis to classify the type of aphasia presented by the patient. It consists of a battery with 28 tasks that assess oral comprehension, oral agility, repetition, naming, oral reading, reading comprehension, and writing. This test was used only for the characterization of the sample, as it is a reference for aphasia diagnosis and for dividing the groups into those with and without predominantly expressive aphasia.

d) Geriatric Depression Scale - GDS-15 (^{Almeida & Almeida, 1999}): a 15-item instrument with yes/no responses, answered only by participants aged 60 and over, to exclude those with moderate to severe depressive symptoms (score of 11 points or more). This scale was applied only to ensure that participants met the study's inclusion criteria.

e) Beck Depression Inventory (BDI, ^{Cunha, 2001}): Completed by adults up to 59 years of age to exclude participants with moderate to severe depressive symptoms. The BDI has 21 items, in which the participant indicates whether they have experienced depressive symptoms in the past few weeks, with scores ranging from zero to four. This instrument was applied only to ensure the study's inclusion criteria, with participants with a score of 20 points or higher being excluded.

f) Mini-Mental State Examination (^{Folstein et al., 1975}): A screening instrument used to check for signs of cognitive decline. This test was only administered to neurologically healthy adults to ensure the inclusion criteria, using cutoff points for education according to ^{Kochhann et al. (2010)}.

The assessments took place in rooms at the research institution or in the participants’ homes (four post-stroke adults) in case of difficulty with transportation. The assessments were conducted in quiet, well-lit environments with no interruptions in both situations. Participants completed the entire assessment in an average of two sessions, each lasting approximately one hour and thirty minutes.

Data analysis

Initially, descriptive analyses of the participants’ performances in the NEUPSILIN-L were conducted. As the data did not follow a normal distribution, non-parametric analyses were performed using Kruskal-Wallis tests, with corrections for multiple comparisons, to compare the three groups (LHD, RHD, and healthy adults) and their scores in the NEUPSILIN-L subtests. Mann-Whitney U tests (non-parametric) were used to compare the groups of patients with and without aphasia. Wilcoxon Signed-Rank tests were used for intra-group task comparisons. Furthermore, reliability was estimated using internal consistency analysis by calculating Cronbach’s alpha and McDonald’s omega for all subtests of the instrument. Spearman correlations were also conducted only with tasks that showed modifications in the instrument adaptation, as the unmodified tasks had already demonstrated reliability in other studies (see ^{Pawlowski et al., 2008}). The statistical significance level for all analyses was set at α < 0.05. Analyses were performed using the SPSS (Statistical Package for the Social Sciences) version 27.0 software.

a) Comparative groups (LHD, RHD, and neurologically healthy adults) in the NEUPSILIN-L

In the comparison between groups, as shown in Table 2, it was identified that patients with LHD had lower scores compared to both the RHD group and healthy adults in tasks related to temporal-spatial orientation (p = 0.010; p < 0.01), time orientation (p = 0.008; p < 0.001), and spatial orientation (p < 0.01; p = 0.015), in both motor response (MR) and oral response (OR). Additionally, the LHD group performed worse than the other groups in tasks related to working memory (p = 0.014; p < 0.001), auditory word span in sentences (p = 0.033; p = 0.001), oral language (p = 0.009; p < 0.001), written language (p = 0.015; p < 0.001), repetition (p = 0.004; p < 0.01), automatic language (p = 0.005; p < 0.001), inference processing (OR) (p = 0.001; p < 0.001), written comprehension (p = 0.034; p < 0.001), spontaneous writing (p = 0.022; p < 0.001), dictated writing (p = 0.036; p < 0.001), and orthographic (p = 0.007; p < 0.001) and semantic (p = 0.039; p < 0.001) verbal fluency. The RHD and healthy groups did not differ in these tasks.

The LHD group had lower scores only compared to the healthy adult group in tasks related to attention (OR and MR; p < 0.001), reverse counting (p < 0.001), digit sequence repetition (p < 0.001), backward digit ordering (p < 0.001), episodic-semantic verbal memory (p = 0.028), immediate recall (p = 0.001), semantic (OR; p = 0.002), visual (p = 0.017), and prospective (p = 0.017) memory, arithmetic skills (p < 0.001), naming (p = 0.001), and ideomotor (p = 0.024), constructive (p = 0.002), and reflective (p = 0.034) praxis. The LHD and RHD groups did not differ in these tasks.

The RHD group performed worse than the healthy adults (p = 0.002) and the LHD group (p = 0.042) in the line perception task and worse than the healthy adults in visual perception (p = 0.038). The hemineglect task reached a ceiling effect, meaning that none of the participants in the sample scored zero.

b) Comparisons between groups with and without aphasia

When dividing the LHD group into adults with predominantly expressive aphasia (n = 16) and those without aphasia (n = 22), statistically significant differences were identified between these groups in tasks related to inference processing by OR (U = 246.000; p = 0.039) and MR (U = 248.000; p = 0.033), oral language by OR (U = 251.500; p = 0.024) and MR (U = 244.000; p = 0.045), and reflexive praxis (U = 247.500; p = 0.033), with the first group performing worse than the second.

c) Comparisons between oral and motor responses in post-stroke patients with LHD

For some subtests, motor responses were provided to the patient during the NEUPSILIN-L assessment. According to the results, motor responses for the subtests of temporal orientation, spatial orientation, semantic memory, and inference processing were significantly better than the oral responses for the same subtests (Table 3).

d) Analysis of internal consistency

The items of the NEUPSILIN-L showed a Cronbach’s alpha and McDonald’s omega of 0.931 and 0.950, respectively. The results of the correlations between the tasks of the NEUPSILIN-L, which underwent modifications compared to the original instrument, are presented in Tables 4 and 5. In the healthy sample, only the semantic memory task (OR) did not show correlations with any task of the NEUPSILIN-L (Table 4). Other subtests showed weak to moderate correlations with each other. In the group of adults with stroke, there were weak, moderate, and strong correlations between all the NEUPSILIN-L tasks (Table 5).

Comparative groups (LHD, RHD, and neurologically healthy adults) in the NEUPSILIN-L

From the analysis of group comparisons, the hypothesis that adults with LHD would perform worse in the NEUPSILIN-L language subtests compared to the other groups was confirmed. Several studies support the findings of language impairments in post-stroke adults (^{Basilakos, 2015}; ^{Boukrina et al., 2015}; ^{Fontoura et al., 2013}; Pawlowski et al., 2013; ^{Rapp, 2011}; ^{Rodrigues et al., 2019}, 2020). ^{Basilakos (2015)} found an association between speech apraxia and aphasia with motor cortical regions, somatosensory areas, the temporal lobe, and the prefrontal cortex of the left cerebral hemisphere. ^{Boukrina et al. (2015)} identified lower performance in patients with LHD in tasks involving semantic, orthographic, and phonological processing. These findings are consistent with the lower performance of participants with LHD in oral language, repetition, automatic language, inference processing, written language, spontaneous writing, and dictated writing tasks of the NEUPSILIN-L when compared to the RHD and healthy groups. Therefore, the relationship between the processing of oral and written language predominantly performed by the left cerebral hemisphere is evident.

The hypothesis that adults with LHD would perform worse in the praxis tasks was also confirmed. Deficits in reading, spontaneous writing, copied writing, dictated writing, and praxis tasks show correlations with left hemisphere functioning, as well as with each other, as they are all related to social interaction through communication (^{Pawlowski et al., 2013b}; ^{Rodrigues et al., 2011}). Accordingly, apraxic deficits and language comprehension impairments are significantly associated with lesions in the left cerebral hemisphere. The NEUPSILIN-L evaluation results also indicated lower performance in the praxis tasks (ideomotor, constructive, and reflective) in the patients with LHD compared to the other groups, showing a relationship between these deficits and the affected hemisphere (^{Rodrigues et al., 2011}).

The results with the NEUPSILIN-L confirmed findings from the original NEUPSILIN when comparing the performance of adults with LHD to neurologically healthy adults (^{Pawlowski et al., 2013b}). Adults with post-stroke LHD showed a significant reduction in performance in language (oral and written), working memory, and ideomotor praxis tasks. The NEUPSILIN-L working memory tasks depend on language (comprehension and expression) to be performed correctly, which justifies the poorer performance of the LHD group.

Another confirmed hypothesis was that adults with RHD would perform worse than healthy adults in the visual perception subtest. Participants with RHD showed poorer performance in the line perception task of the NEUPSILIN-L. ^{Russell et al. (2013)} demonstrated that RHD patients had a reduced effective visual field, mainly for identifying items to their left, indicating the importance of the right hemisphere in visual perception processing. ^{Vossel et al. (2013)} also showed the emergence of deficits such as hemineglect and agnosia, directly impacting the daily living activities of these patients. Cerebral areas related to this type of deficit were the frontal cortex, cingulate cortex, and white matter of the right hemisphere.

The other visual perception tasks (face perception and recognition), as well as the hemineglect task of the NEUPSILIN-L, did not differentiate the three groups, although it is common for RHD patients to present visual hemineglect after a stroke (^{Nurmi et al., 2010}). It is likely that the small number of items in the instrument, along with the small sample size, compromised the significance of the differences. Additionally, the hemineglect task may have been too easy to execute, not being sensitive to patients with chronic cerebrovascular lesions or more subtle deficits.

The hypotheses that participants with RHD would perform worse in the inference processing and constructive praxis tasks when compared to the healthy group were not confirmed. This fact may be related to the small sample size and the limited variability in the NEUPSILIN-L scores, which were not sensitive enough to detect these difficulties. Qualitative analyses of error types, such as observing whether the drawing suggests hemineglect and whether there are concrete responses in metaphor interpretation, may complement the quantitative analysis of the NEUPSILIN-L and contribute to the understanding of deficit processing.

In general, the results of the comparisons between the three groups (LHD, RHD, and healthy) using the NEUPSILIN-L corroborated findings already documented in the literature on cerebral hemisphere specialization (^{Rapp, 2011}; ^{Rodrigues et al., 2013}; ^{Tynterova, Perepelitsa, & Golubev, 2022}). Furthermore, the instrument was able to help understand the neuropsychological profile of patients with LHD, RHD, and neurologically healthy adults. Therefore, it can be said that the NEUPSILIN-L presents evidence of criterion-related validity by demonstrating differences in performance according to the participants’ neurological clinical status. However, the study limitations include the small number of participants and the heterogeneity of cerebrovascular lesions, which may not have shown statistically significant differences. Analyses with larger and more homogeneous groups are recommended for future studies.

b) Comparisons between groups with and without aphasia

The hypothesis that participants with LHD and aphasia would have lower scores on the NEUPSILIN-L was also corroborated, again demonstrating criterion-related validity evidence. Differences in oral language skills between groups with and without expressive aphasia are expected, as these tasks involve naming, comprehension, automatic speech, and repetition, which may be impaired in these cases (Fontoura et al., 2014). The task of processing inferences is also speech-dependent, which may have hindered the performance of patients with aphasia. On the other hand, reflexive praxis is not a language-dependent task for execution, although participants must understand its instructions. Therefore, difficulties in attention and comprehension of what should be done may have highlighted these differences in this task.

It is important to note that participants in the groups with and without aphasia, despite having left hemisphere lesions, are heterogeneous in terms of the extent of the lesion, affected lobes, and time post-stroke, which are variables that influence neuropsychological deficits (^{Rodrigues et al., 2011}; 2018). This constitutes a further limitation of the present study. Patients with aphasia may have a greater number of neuropsychological deficits associated with extensive brain lesions, affecting various cognitive abilities such as working memory and attention (^{Crivelli et al., 2023}; ^{Lee & Pyun, 2014}). Furthermore, even if patients have lesions in the same specific brain areas, they may manifest different impairments. It is suggested that case series studies with the same type of brain lesion or studies with homogeneous groups, using the NEUPSILIN-L, could contribute to research in the field.

c) Comparisons between oral and motor responses in post-stroke patients with LHD

Participants in the clinical groups showed better performance in the NEUPSILIN-L in tasks with motor response options compared to oral responses in the orientation, memory, and inferential processing subtests. It is possible that having response options facilitates neuropsychological assessment, as participants can recognize the correct alternative instead of recalling it. In the application of the NEUPSILIN-L, to standardize the instrument, all participants initially provided oral responses, followed by motor responses. This inclusion of motor responses has proven valuable in clinical practice, offering greater sensitivity when compared to tests relying solely on oral responses, especially for patients with aphasia (^{Crivelli et al., 2023}). Furthermore, having visual options to point to for the answer may provide a more reliable assessment of the specific neuropsychological function in question. This result suggests that the motor modality, an innovation in the NEUPSILIN-L, allows the patient to respond to the subtest without the effect of expressive or oral language impairments affecting their score.

On the other hand, having alternatives for the participant to point to may make the assessment easier, and the participant may “guess” the answer, leading to a false-negative result (not demonstrating a deficit in the neuropsychological function). Therefore, it is recommended that the evaluator pays attention to the accuracy of the responses, hesitations, and test duration, and analyzes the patient’s results in the NEUPSILIN-L to assess their neuropsychological profile. Despite these limitations, the use of the instrument is recommended for patients with expressive language difficulties, such as aphasia, mutism, and communication disorders, aiming to circumvent the language deficit and assess the target neuropsychological construct in clinical contexts.

d) Analysis of internal consistency

Finally, internal consistency analyses demonstrated results above 0.90, with alphas between 0.60 to 0.70 being considered borderline for an acceptable result (Hair, Black, Babin, Anderson, & Tatham, 2009). Therefore, the results of the NEUPSILIN-L demonstrate high internal consistency. Additionally, in general, the neuropsychological tasks showed statistically significant weak, moderate, and strong correlations in both groups, indicating that they assess related functions and demonstrate interdependence (^{Rodrigues et al., 2019}). Moreover, many tasks are language-dependent, which supports the correlations found between the subtests. From the correlations, it can be observed that the motor response alternatives continue to assess the main neuropsychological functions. Therefore, the results of the NEUPSILIN-L demonstrate high reliability. The weakness of the correlations among healthy individuals may have been due to this group exhibiting a ceiling effect in many of the tasks of this instrument, compromising the significance between them. In contrast, in the group of adults with stroke, there were weak, moderate, and strong correlations between all the NEUPSILIN-L tasks. The greater variability in scores in this group, compared to the healthy individuals, may have favored the stronger and more significant correlations.