Cognitive reserve (CR) refers to “the adaptability of cognitive processes that helps to explain differential susceptibility of cognitive abilities or day-to-day function to brain aging, pathology, or insult” (Stern et al., 2020). This concept helps to explain individual differences in cognitive or functional decline related to aging or brain disease (Stern et al., 2020). It has been suggested that CR is related to compensatory mechanisms (Gonzalez-Burgos et al., 2020), partially related to age differences in brain activation (Festini et al., 2018).

Research on CR has primarily focused on verbal proxies, which are presumably related to the left hemisphere of the brain (Springer et al., 2005; Bartrés-Faz et al., 2009; Solé-Padullés et al., 2009; Bosch et al., 2010; Steffener et al., 2011). Thus, research on the right hemisphere and its cognitive functions has been limited. However, the “right hemi-aging model” (RHAM) suggests that the right hemisphere is more vulnerable to aging than the left hemisphere (Dolcos et al., 2002).

Research on the right hemisphere and its functions has increased over the last few decades. In a recent systematic review (Cocquyt et al., 2017), it was suggested that the right hemisphere could have a role in compensatory processes, such as early spontaneous language recovery in aphasia. In addition, some authors have suggested the existence of a CR network within the right hemisphere (Robertson, 2014). Brosnan et al. (2018) found an association between a high CR and an increased participation of the right hemisphere in visual processing. Similarly, Fleck et al. (2017) highlighted the role of the right hemisphere in people showing better global brain functioning and high CR.

Among proxies of CR, different measures of premorbid intelligence quotient (IQ) have been widely used to operationalize CR (Stern et al., 2020). These premorbid IQ proxies typically include several subtests of the Wechsler Adult Intelligence Scale (WAIS, Wechsler, 1997), including the Information subtest (Ferreira et al., 2017; Gonzalez-Burgos et al., 2020), the Vocabulary subtest (Solé-Padullés et al., 2009), the Similarity subtest and the Digit Span subtest (Colombo et al., 2018). For instance, the Information subtest is commonly considered as a measure of crystallized intelligence (Wechsler, 2008). In the cohort used in the present study, the Information subtest showed the greatest mediation capacity for the effect of cortical thinning on cognition, as compared with the Vocabulary subtest and other proxies of CR (Ferreira et al., 2016). Furthermore, the Information subtest better represented achievements and/or use of educational opportunities versus educational attainment or years of education (Correia et al., 2015).

Hence, with regards to premorbid IQ proxies of CR, measures presumably related to the left hemisphere have also been used more. To the best of the authors’ knowledge, how non-verbal or visual (right-hemisphere) premorbid IQ proxies relate to CR has not been studied. The Block design subtest is an IQ measure of the WAIS that assesses visual domains such as visuoperceptive, visuospatial and visuoconstructive functions (Wechsler, 1997). The Block design subtest is one of the most sensitive tests to assess brain damage (Lezak et al., 2012) and it has a fluid component strongly influenced by age (Wechsler, 2008). Therefore, the Block design subtest is considered as a measure of fluid intelligence (Wechsler, 2008). As opposed to crystallized intelligence, fluid intelligence refers to the capacity for logical reasoning and the ability to solve novel problems regardless of previously acquired knowledge, typically assessed through non-verbal cognitive tests (Cattell, 1971; Yuan et al., 2018). Furthermore, the Block design subtest correlates with CR proxies such as years of education, WAIS-III Vocabulary and Information subtests and CR questionnaires (Ferreira et al., 2016).

The main objective of this study was to determine the potential role of the Block design subtest as a possible non-verbal and presumably right-hemisphere-related proxy of CR. The similarity between the Block design subtest performance and the Information subtest performance was tested, the latter being a widely used verbal IQ proxy of CR. Firstly, the Block design association with CR proxies needs to be demonstrated. In order to do this, the association between the Block design and Information subtests was tested, as well as the association of both CR proxies with an established questionnaire of CR. Secondly, the fact that the Block design and Information subtests are similarly associated with performance in different cognitive domains needs to be demonstrated. Thirdly, the capacity of Block design to minimize the negative effect of older age over cognitive performance on cross-sectional data was tested; in this case, this subtest should be considered as a CR proxy. This effect should then be compared to that of the Information subtest. The main hypothesis of the present study is that the referred Block design subtest would perform similarly to the Information subtest, demonstrating its capacity as a proxy of CR. This could have clinical implications for neuropsychological assessment and interpretation of test results, providing a non-language-dependent alternative to traditional verbal CR proxies.

The demographic characteristics of the cohort (N = 391) are shown in Table 2. There was a strong inverse association between the raw Block design subtest and age (r = −0.572; p < 0.001; Figure 1A). After applying the correction from Amato et al. (2006), the new Pearson correlation between the corrected Block design subtest and age demonstrated that the variance of age on Block design was completely removed, as expected (r = 0; p = 1.0; Figure 1B).

The present study aimed to determine the potential role of the Block design subtest as a possible non-verbal proxy of CR. The results show that the corrected Block design subtest was strongly correlated with a questionnaire of CR and with the Information subtest, a widely used verbal premorbid IQ proxy of CR. In addition, higher scores in the corrected Block design subtest minimized the negative effect of an older age on cognitive performance in exactly the same cognitive domains as the Information subtest, thus behaving as a CR proxy. The associations of the corrected Block design subtest with cognitive domains were also similar to those of the Information subtest, thus further supporting the similarities between the Block design subtest and the verbal CR proxy of the Information subtest.

The significant correlation between the Block design and Information subtests is consistent with previous findings (Ferreira et al., 2016), where the same cohort was used but the analyses included individuals between 38 and 52 years of age. These previous findings showed that the Block design subtest correlated with cognitive reserve proxies such as WAIS-III Vocabulary and Information subtests, years of education, and the CRQ. Also in the same cohort (Correia et al., 2015) reported that participants between 60 and 80 years of age who performed higher on the Information subtest performed better on the Block design subtest. The present study, thus, managed to generalize the findings from these two previous narrow-age-range studies to a wider age range, from 38 to 80 years of age. This is above all a first piece of evidence in favor of considering the Block design subtest as a potential non-verbal proxy of CR.

Another result of the present study is the significant association of the Block design (and Information subtest) with cognitive performance. In particular, the corrected Block design subtests showed a significant association with the verbal memory domain, the visual memory and visuospatial domain, the executive and premotor functions domain and the processing speed domain, in a such a way that higher scores in Block design were associated with a higher cognitive performance. The association between the Block design subtest and the verbal memory domain may be explained by the strong component of the executive functions of both tasks. It is possible that shared processes such as working memory, planning, developing strategies and processing speed apply in both the Block design and the verbal memory domain. In fact, some authors have reported an important association between verbal memory and executive functions (Duff et al., 2005). The association of Block design with the visual memory and visuospatial domain is reasonable because of the involvement of visuoperceptual, visuospatial and visuoconstructive functions in the Block design performance. The association between the corrected Block design subtest and executive and premotor functions could be interpreted by the participation of premotor functions in the integration of motor skills and learned action sequences (Damasio et al., 2011), which may be involved in Block design performance. Finally, the association between Block design and processing speed is not surprising since the Block design subtest is a timed test with a clear processing speed component (Lezak et al., 2012). Previous studies showed that processing speed predicted a substantial part of the variance in the Block design subtest (Bugg et al., 2006; Schubert et al., 2019).

The associations between the Information subtest and cognitive performance are reminiscent of the associations described above for the corrected Block design subtest. Briefly, the results for the verbal memory domain, visual memory-visuospatial domain, the executive-premotor domain and processing speed domain are consistent with previous studies from the same cohort using different age ranges (Correia et al., 2015; Ferreira et al., 2016). Correia et al. (2015) found a significant effect of the Information subtest scores on verbal memory, visual memory, visuospatial, visuoconstructive, executive and premotor and processing speed tasks. Similarly, Ferreira et al. (2016) found a significant correlation of the Information subtest with verbal memory, visuospatial, visuoconstructive and executive tasks. Taken together, the similar associations obtained for the Block design and Information subtests with cognitive performance serve as another piece of evidence in favor of Block design as a potential non-verbal proxy of CR.

It should be mentioned that the significant interactions between the corrected Block design subtest and age on cognitive performance revealed that higher scores in the corrected Block design subtest minimized the negative effect of age on cognition, in a similar way to the results obtained for the Information subtest. These results are consistent with the predictions of CR theory in normal aging. For example, different studies have demonstrated a slower rate of age-related cognitive decline in individuals with a higher CR (Zahodne et al., 2015). The findings from the interaction of the Information subtest with age in the prediction of cognitive domains enabled the authors to characterize the traditional verbal CR proxy of Information and establish it as a reference for comparison with the proposed non-verbal CR proxy of the corrected Block design.

This study has some limitations. The correlation analysis for the corrected Block design and Information subtests with the CRQ was based on a small subsample of sixty-seven participants whose data were available from a previous study (Ferreira et al., 2016). However, for the purpose of that specific analysis, the statistical power was sufficient to capture the statistically significant association of the corrected Block design subtest with a stablished measure of CR (i.e., the CRQ), which was adequate for the aim of the present study. The inclusion of the CRQ variable made it possible to meet the first aim of the study, and it was not used in further analyses, hence not limiting the number of participants for the analyses in the second and third aims. Another limitation is that potential gender-related differences in cognitive performance and cognitive reserve were not explored. However, to minimize this problem gender differences from the design of this study were controlled for (by setting up a cohort balanced in gender). Schooling years is a common proxy of CR. In this study, the corrected Block design subtest was validated against the Information subtest and the CRQ. These are two well-established proxies of CR and, in particular, the Information subtest performs better than schooling years as a CR proxy in the cohort (Correia et al., 2015; Ferreira et al., 2016). Expanding the present validation here to compare the corrected Block design subtest against schooling years in other different cohorts may be an interesting prospect for future studies. Furthermore, the Block design subtest has both fluid and crystallized components. While the premorbid crystallized component of cognition usually remains stable with aging and in pathology, reflecting fundamental features of CR, the fluid component changes with age. Therefore, the formula proposed by Amato et al. (2006) to statistically remove the variance associated with age in Block design was applied, as a reasonable attempt to remove the fluid component. Although the authors are aware that this is an indirect correction method, there are no strategies in previous research to reduce the fluid component of a given test, to the best of the authors’ knowledge. An alternative strategy could be to isolate the crystallized component of a given test by capturing the common variance between the test (Block design in the present study) and crystallized measures such as the Information subtest. However, such a strategy would completely clash with the design of the present study. Another potential limitation is that the authors accepted the assumption that verbal tests are more related to the functioning of the left hemisphere while non-verbal tests are more related to the functioning of the right hemisphere. Although this assumption is widely accepted in the field, it is almost impossible to definitively prove it. A common strategy to increase the chances that verbal abilities are primarily driven by left hemisphere and non-verbal abilities are primarily driven by right hemisphere is to limit the study population to only include right-handed individuals. The evidence behind this strategy is that around 90% of right-handed individuals have verbal abilities mostly implemented in the left hemisphere (Papadatou-Pastou, 2011). Hence, only right-handed individuals were included in the present study. Furthermore, the study was approached in terms of verbal versus non-verbal proxies, trying to avoid references to brain hemispheric specialization. This is still a valid and important approach because a non-verbal proxy of CR provides the opportunity to estimate CR in populations with low educational levels or with compromised language functions (irrespective of hemispheric specialization). Finally, the Block design subtest was tested as a potential proxy of CR in cognitively unimpaired people. This was necessary to establish the findings in the normal population. Since CR was assessed in the context of increasing age, future studies should expand the present study to demonstrate the potential of Block design as a proxy of CR in individuals with a brain pathology.

In conclusion, it has been demonstrated here that the corrected Block design subtest is significantly associated with two established proxies of CR, it correlates with cognitive performance and high scores in Block design have the capacity to minimize the negative effect of an older age on cognitive performance. Therefore, the results of this study suggest that the corrected Block design subtest could be used as a non-verbal proxy of CR. This could have several clinical implications such as the possibility to estimate CR in patients with verbal impairments or low levels of education.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by CEIBA; local ethics committee of the University of La Laguna (Spain). The patients/participants provided their written informed consent to participate in this study.

GC–B contributed to organize the database, performed the statistical analyses, and wrote the first draft of the manuscript. RY-P contributed to the data collection and wrote the first draft of the manuscript. YM contributed to the data collection and to the design of the study, supervised the statistical analyses, participated in the interpretation of the results, supervised the writing of the manuscript, wrote sections of the manuscript, and revised the final version of the manuscript. NC contributed to the data collection, supervised the statistical analysis, and revised the final version of the manuscript. JB contributed to the conception and design of the study, obtained funding, co-supervised the study, and revised the final version of the manuscript. DF contributed to the conception and design of the study, wrote sections of the manuscript, contributed to the interpretation of the results, obtained funding, and supervised the study. All authors have read and approved the submitted version of the manuscript.

This research was funded by the Estrategia de Especialización Inteligente de Canarias RIS3 de la Consejería de Economía, Industria, Comercio y Conocimiento del Gobierno de Canarias, co-funded by the Programa Operativo FEDER Canarias 2014–2020 (ProID2020010063), and Universidad Fernando Pessoa Canarias. DF receives funding from the Center for Innovative Medicine (CIMED, 20200505), the regional agreement on medical training and clinical research (ALF, FoUI-962240) between Stockholm County Council and Karolinska Institutet, Hjärnfonden (FO20200066), Alzheimerfonden (AF-968032), Demensfonden, Neurofonden, Stiftelsen För Gamla Tjänarinnor, Karolinska Institutet funding for Research, Karolinska Institutet funding for Geriatric Diseases and Gun och Bertil Stohnes Stiftelse.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.