Dementia is a highly prevalent disorder, with considerable social and economic costs, particularly in developing countries (Nitrini et al, 2009; Prince et al, 2013). Studying modifiable risk factors associated with dementia is important for the potential development of new strategies that may reverse the increasing incidence of this disorder. Illiteracy is one of the many modifiable risk factors for dementia (Norton et al, 2014; Valenzuela and Sachdev, 2006). The mechanisms underlying the protection provided by higher literacy levels seem to be related to the influence of literacy on cognition, brain function, and brain structure.
According to recent studies, literacy strongly influences cognition (Arenaza-Urquijo et al, 2013; Foubert-Samier et al, 2012; Opdebeeck et al, 2016), and literate individuals perform better on verbal (Nitrini et al, 2004) and even nonverbal tasks (Nitrini et al, 2005) than their illiterate counterparts. Positron emission tomography and functional magnetic resonance studies have shown the influence of literacy levels on brain function by demonstrating higher cerebral perfusion in parietal and occipital regions during verbal tasks in literate versus illiterate participants (Castro-Caldas et al, 1998; Dehaene et al, 2010, 2015). These findings suggest that literacy can affect specific patterns of brain activation related to cognitive tasks. Regarding brain structure, voxel-based morphometry studies have demonstrated higher gray and white matter intensity in literate versus illiterate participants (Carreiras et al, 2009; Petersson et al, 2007). Such findings suggest the role of literacy in shaping the inherent cerebral anatomic organization and, perhaps, in compensating for the effects of amyloid deposition in people with Alzheimer disease (Arenaza-Urquijo et al, 2017).
Relatively recent research has focused on examining aspects of white matter microstructure using advanced magnetic resonance imaging (MRI) techniques such as diffusion tensor imaging (DTI) (Ciccarelli et al, 2001; Hagmann et al, 2006). DTI can capture the interference in the movement of water molecules caused by distinctive properties of components of the brain’s white matter. Common DTI measures include fractional anisotropy (FA) and mean diffusivity (MD), among others. FA refers to the longitudinal directionality of the water molecules’ movement, and MD represents the radial directionality of the water molecules’ movement. FA and MD are believed to be sensitive to axonal ordering, axonal density, and the degree of myelination (Jones et al, 2013). Higher FA values often correlate with higher white matter microstructure integrity and organization, whereas higher MD values usually correlate with lower white matter microstructure integrity (Beaulieu, 2002). FA and MD values indirectly reflect the integrity of white matter microstructure and organization in the brain and have been used as surrogates for measures of brain structural connectivity, which is a key factor in supporting cognitive functioning (Johansen-Berg, 2010; Kanai and Rees, 2011; Mesulam, 2005).
Literacy acquisition enhances the connections between a person’s visual and language systems, changing his or her ventral visual pathway, promoting lateralization, and ultimately affecting visual processing and phonological awareness (Dehaene et al, 2015). Literacy has also been linked to measures of white matter microstructure. For example, bilingual people have been found to have higher FA values across many white matter tracts than monolingual people (Pliatsikas et al, 2015). In addition, learning to read and write has been associated with an increase in FA values in the corpus callosum among former functional illiterates (Boltzmann et al, 2017). Furthermore, Thiebaut de Schotten et al (2014) demonstrated through tractography analysis that middle-aged literate participants with 7 years of education had higher FA values in the left arcuate fasciculus than did their illiterate counterparts. Finally, cognitive training in older adults is associated with less decline in FA over time (Cao et al, 2016), and highly educated older adults seem to overcome the aging-related decline in FA values better than less educated older adults (Vaqué-Alcázar et al, 2017). However, the role of low literacy (defined as <4 years of formal education) versus illiteracy (defined as no formal education) has been underexplored in the literature. Because white matter in the brain is vulnerable to disease and susceptible to environmental changes, studying the structure of white matter is paramount in understanding sensitive mechanisms related to the interface between literacy and brain structure.
Dementia is highly prevalent in low- and middle-income countries where the educational level among the elderly is low (United Nations Educational, Scientific and Cultural Organization Institute for Statistics, 2016). To determine whether illiteracy is a modifiable risk factor for dementia, a study comparing elderly participants with low levels of literacy versus complete illiteracy is important. The relevance of this field of research is not limited to developing countries but includes industrialized nations experiencing an increasing number of immigrant populations with a low-literacy background. Thus, discovering the potential benefits of low literacy levels on brain structure may contribute to the development of strategies to prevent dementia. One such strategy in both low- and high-income countries might be to raise the literacy level of illiterate older adults who are at a high risk of developing dementia.
We conducted a study using DTI to compare low-literate with illiterate elderly Brazilians in order to identify differences in the integrity of their white matter microstructure. We hypothesized that the low-literate group would have higher FA and lower MD values than the illiterate group, ie, higher white matter integrity. This finding would suggest a positive association between some degree of literacy and brain structural connectivity.
We selected participants for this research from the Pietà study, a community-based investigation on brain aging conducted in Caeté (Minas Gerais state), Brazil (Caramelli et al, 2011). The aim of the Pietà study was to evaluate the oldest-old adults living in the city to identify aspects related to healthy brain aging such as cardiovascular diseases, health habits (smoking, alcohol abuse, exercise, leisure activities, diet), cognition, and inflammatory markers. Drawing on census records, investigators invited all persons 75 years old or older (n=1251) to participate in the research, of whom 51.1% (n=639) were evaluated.
The Pietà study’s participants were evaluated by a group of experienced neurologists and geriatricians, along with a psychiatrist. The participants underwent a detailed neurologic and neuropsychological examination to determine their clinical, functional, neurologic, and psychiatric status. For the purpose of the Pietà study, the participants were classified as cognitively healthy, cognitive impairment-no dementia (CIND) (Ebly et al, 1995; Graham et al, 1997), or dementia. Further details on the characterization of this sample are provided elsewhere (Caramelli et al, 2011).
For the Pietà study, participants were considered illiterate if they had never received any formal education and were unable to perform the following tasks from the Mini-Mental State Examination (MMSE) (Brucki et al, 2003; Folstein et al, 1975): to read the phrase “Close your eyes” and to write a sentence. Participants were considered literate if they had attended school for at least 1 year and were able to read the same phrase from the MMSE and to write a sentence.
The study used a questionnaire from the Brazilian Association for Market Research Institutes scale (Associação Brasileira de Empresas de Pesquisa [Brazilian Association of Research Companies], 2003) to assess each participant’s socioeconomic level. The questionnaire considers the presence and quantity of domestic appliances, bathrooms, and automobiles at home; the householder’s educational level; and the presence of a paid housekeeper. The scale ranges from A=high socioeconomic level to E=very low socioeconomic level (Associação Brasileira de Empresas de Pesquisa [Brazilian Association of Research Companies], 2003).
The Pietà study’s participants were recruited in 2008. In 2009, 200 random participants were invited to be reassessed and to undergo MRI scans. Of those, 11 declined to participate, and a total of 189 MRI scans were acquired (Figure 1).
The Pietà study was approved by the ethics committee of the Federal University of Minas Gerais. All participants or their legally authorized representatives provided written informed consent.
Neuroimaging Data Acquisition and Processing
The MRI scans were acquired in a 3-T Philips Achieve scanner. Of the participants who underwent MRI scans (n=189), 49 had the DTI acquisition (Resende et al, 2017). In most of the scans, the DTI sequence was not acquired because of a lack of time to perform a longer sequence or because the participant was noncooperative.
The diffusion-weighted images were obtained in the axial plane with single-shot, spin-echo echo-planar sequences in the axial plane (repetition time=8000 msec, echo time=65 msec, field of view=240 mm, matrix=96×96 [reconstructed 128×128], slice thickness=2.0 mm with a 1.99-mm gap between slices). Diffusion sensitization gradients were applied in 32 noncollinear directions, with a b factor of 1000 seconds/mm2. An experienced radiologist blind to the clinical diagnoses calculated the Fazekas score (Wahlund et al, 2001). The Fazekas score is a visual scale for scoring the severity of white matter hyperintensities on the fluid-attenuated inversion recovery MRI sequence. It ranges from 0 to 3 (0=absence of lesions, 1=focal lesions, 2=beginning confluence of lesions, and 3=diffuse and confluent lesions). The final score ranges from 0 to 6 because it consists of the sum of the scores in the white matter and in the basal ganglia (Wahlund et al, 2001). The same sample of participants who had undergone an MRI evaluation with these same acquisition parameters was previously described in a study that examined the neural correlates of episodic memory in low-literate older adults (Resende et al, 2017).
We visually inspected the diffusion images for artifacts. Eighteen participants had to be excluded at this point because their images had major artifacts. Therefore, our final sample consisted of 31 participants (Figure 1), none of whom had dementia. To correct for subtle movement artifacts and eddy current distortion effects, we coregistered the nondiffusion and diffusion images on the echo-planar imaging readout (Woods et al, 1998). Multivariate fitting and diagonalization (FSL 4.0 FMRIB software) were used to calculate the diffusion tensor for each voxel (Jiang et al, 2006; Pajevic and Pierpaoli, 1999). Constrained nonlinear registration (Image Registration Toolkit) (Pajevic and Pierpaoli, 1999) was used to brain-extract the FA and MD images (Brain Extraction Tool, FSL) and register them to the Montreal Neurological Institute space standard (MNI152). The derived FA and MD data were further analyzed using a whole-brain voxelwise analysis, ie, tract-based spatial statistics (TBSS) (Smith et al, 2006) and a priori analyses of regions of interest (ROIs). These procedures are described in detail elsewhere (Tovar-Moll et al, 2014).
Whole-Brain Voxelwise Analysis
To contrast low-literate with illiterate participants, the t test was carried out in the whole-brain voxelwise analyses of FA and MD using TBSS (FMRIB Software Library, FSL) (Smith et al, 2004), controlling for age. To correct for multiple comparisons, we used the family-wise error rate considering permutation-based nonparametric inference with 10,000 random permutations (FSL Randomise tool) (Nichols and Holmes, 2002; Rueckert et al, 1999) on each voxel of the resulting “mean skeletonized” data (Jiang et al, 2006). After this correction, the results were considered significant at P<0.05, with a reasonable low probability of false positives (Bennett et al, 2009). At a cluster level of correction, we used the cluster-based threshold-free cluster enhancement (Smith and Nichols, 2009). The threshold-skeletonized resulting image was thickened for better visualization.
We used a DTI-MRI atlas of human white matter to determine fiber tract orientation and the ROIs (Hua et al, 2008; Jiang et al, 2006), which were then confirmed by an experienced investigator (F.M.F.). The Johns Hopkins University white matter atlas (Hua et al, 2008) was our reference for the anatomic labels. We placed the ROIs in the inferior longitudinal fasciculus, the inferior fronto-occipital fasciculus, and the superior longitudinal fasciculus because these three fasciculi connect the visual cortex with language areas and may be important for literacy acquisition (Catani and Thiebaut de Schotten, 2012; Dehaene et al, 2010, 2015). A t test was carried out in these regions as described for whole-brain voxelwise analyses using the same corrections for multiple comparisons analyses.
Statistical Analyses of Demographic and Clinical Variables
We used IBM SPSS Statistics 16.0 to perform the statistical analyses. The Kolmogorov-Smirnov test was carried out to assess the parametric distribution. The t test was used for comparing the differences among the groups regarding age and years of education as well as MMSE and Fazekas scores, all of which had a parametric distribution. We used the Fisher exact test to investigate differences between categorical variables. The level of significance (α) was set at 0.05, two-tailed, for those statistical tests.
The final sample consisted of 31 participants, 21 of whom were literate (mean years of education=3.4±1.4) and 10 illiterate. In the literate group, the mean age was 79.8±3.8 years; 14 were women and seven were men; 17 were cognitively healthy, and four had CIND. In the illiterate group, the mean age was 80.7±4.1; eight were women and two were men; six were cognitively healthy, and four had CIND. None of the 31 participants had dementia. There were no differences between the two groups in terms of age, sex, or proportion of CIND diagnosis (Table 1).
The literate participants had significantly higher MMSE scores than the illiterate participants, which was an expected difference because scores are highly influenced by test participants’ educational level (Brucki et al, 2003). The two groups had a similar burden of white matter hyperintensities, as assessed through the Fazekas scale (Table 1). Per the socioeconomic questionnaire we applied, the proportion of lower socioeconomic level was higher in the illiterate group. Fifty percent of illiterate participants were in levels D and E, whereas 20% of literate participants were in these low levels, although this difference was not significant (P=0.341) (Table 1). Moreover, no participant was in socioeconomic level A, and only one participant (illiterate) was in level E. Therefore, the entire sample was socioeconomically comparable.
A direct comparison between the literate and illiterate groups using a whole-brain voxelwise TBSS approach did not show a significant difference between the two groups, although there was a tendency for higher FA values in the literate participants (P=0.087). Using the ROI approach, we identified significantly higher FA values in the right superior longitudinal fasciculus (P=0.022) (Figure 2) of the literate participants as well as a tendency for higher FA values in the left superior longitudinal fasciculus (P=0.056) and in the left inferior longitudinal fasciculus (P=0.053) compared with those in the illiterate individuals. There were no significant differences in the MD analysis.
In comparing low-literate and illiterate elderly adults, we found significantly higher FA values in the right superior longitudinal fasciculus in the low-literate group. There was also a tendency toward higher FA values in the left superior longitudinal and left inferior fasciculus in the low-literate group. The mean years of schooling in the low-literate group was 3.4±1.4 (range, 1 to 7), which barely corresponds to primary school according to the International Standard Classification of Education (United Nations Educational, Scientific and Cultural Organization Institute for Statistics, 2012). Because FA values are an indirect measure of white matter microstructure integrity reflecting axonal organization, myelination, and cell membrane integrity, our results suggest that low literacy at the level of primary school might be associated with those white matter properties. However, the specific property and directionality of the effect could not be identified in the current study. In light of the similar, low socioeconomic level of the two groups, the differences found are unlikely to be attributable to nutritional or other socioeconomic factors. Although our results are preliminary because of the small sample size and low statistical power, we consider these findings to be important pilot observations useful to further research aimed at understanding the underlying mechanisms of cognitive reserve in low-literate individuals.
Literacy acquisition is believed to enhance the connections between the visual and language systems, ie, the occipital and temporal lobes of the brain (Dehaene et al, 2015). Therefore, white matter bundles that connect those regions, such as the superior longitudinal fasciculus, the inferior longitudinal fasciculus, and the inferior fronto-occipital fasciculus, are potential targets of changes associated with literacy acquisition. Our significant finding involved the superior longitudinal fasciculus on the right side. This seems counterintuitive because language is generally left lateralized, and literacy acquisition is highly dependent on language abilities. However, hemispheric lateralization is influenced by an individual’s educational level. Lecours et al (1988) found that when comparing patients with a right-sided stroke, the ones who were illiterate had more difficulties in naming tasks than the literate ones. Furthermore, Petersson et al (2007) demonstrated that being illiterate manifested as more right-lateralized than left-lateralized brain correlates in language tasks. If we had had a larger sample, we might have observed significant differences in the left hemisphere of the participants, shifting the association from right-sided to bilateral, in accordance with bilateralism. Bilateralism has been described previously in the literature regarding memory and language correlates in the elderly (Buckner, 2004).
The neurobiological basis of cognitive reserve in the elderly is a research topic of utmost importance because it has the potential to reveal protective factors against dementia. Hence, the comparison between elderly participants with basic literacy levels and those who are unable to read and write brings to light the ability to explore the relationship between aging, brain structural connectivity, and vulnerability.
Recent findings have suggested that age-associated changes in the brain’s structure depend on active stimulation of the brain, and literacy plays an important role in modulating cognitive reserve (Fratiglioni and Wang, 2007). Cognitive reserve, in turn, is thought to be an important factor for successful brain aging (Buckner, 2004). Previous studies have found that higher literacy levels counteract the effect of tau (Hoenig et al, 2017), amyloid (Arenaza-Urquijo et al, 2017), and vascular (Farfel et al, 2013) pathology in the brain. In other words, high-literate individuals need more pathology to manifest symptoms than do low-literate individuals. However, those studies did not reveal the mechanism by which literacy counteracted the effect of those pathological changes.
It has been proposed that literacy acquisition promotes the development of compensatory strategies either by increasing the connections between brain regions or by having more efficient connections that are more resilient to brain pathology (Stern, 2012). To corroborate this theory, studies have shown that high-literate adults have more efficient patterns of brain activation during cognitive tasks than do low-literate adults (Bartrés-Faz et al, 2009; Scarmeas et al, 2003). Furthermore, increased values of gray and white matter intensity (Boltzmann et al, 2017) and functional connectivity (Skeide et al, 2017) in low-literate adults (found after training to improve their reading skills) suggest that literacy acquisition might have an effect on building more brain volume and connections to buffer cognitive decline when the brain is affected by a neurodegenerative disease. In view of the observation of neurogenesis in adult mice exposed to a cognitively enriching environment (Kempermann et al, 2002), another possible explanation is that neurogenesis may also take place in humans after cognitive stimuli.
Our study has notable limitations. The sample size was small because we could not perform DTI on all of the individuals who underwent an MRI, and we had many images with artifacts. Selection bias is possible because more independent and cognitively intact older adults may have been more easily transported to the MRI site and better able to remain still inside the scanner. The limited sample size resulted in an underpowered sample that restricted further conclusions and the generalizability of the findings.
Considering the potential problems with the sampling, we conservatively analyzed the images using rigorous multiple comparison corrections at the voxel level, using a high magnetic field (3-T MRI) and the TBSS approach, which together make the interpretations of the findings more reliable (Lavdas et al, 2014; Smith et al, 2006). Because we corrected for multiple comparisons at the voxel level, we did not apply a post hoc correction on top of the ROI contrast analyses, an approach adopted in previous studies in the field (Fontenelle et al, 2011; Kurumaji et al, 2017; Park et al, 2018). Furthermore, the two groups had a similar socioeconomic level, but we did not include this variable in the analyses. Because low literacy levels are linked to low socioeconomic status, which is associated with factors such as poor nutrition and lack of health care access, we acknowledge that it is difficult to disentangle whether our results were also associated with these other variables. Although preliminary, this is the first study in older adults to contrast white matter microstructure in illiterate individuals with that in low-literate individuals, with the conclusion that any degree of literacy might be better for brain health than none at all.
Future analyses of a broader sample of participants are necessary to confirm the results of this study. Furthermore, studies including different literacy levels and wider ranges of age and cognitive performance will clarify the specifics of the association between white matter microstructure and literacy. Clearly, longitudinal studies will be needed to investigate the potential of preventing cognitive decline by providing basic literacy to older adults.
The authors thank the study participants, the Pietà study group, and the D’Or Institute for Research and Education. The authors especially thank Luciana Costa Silva for analysis of the Fazekas score in all of the images, Alissa Bernstein for the English review, and Laboratório Hermes Pardini for the MRI acquisition support.
Arenaza-Urquijo EM, Bejanin A, Gonneaud J, et al. 2017. Association between educational attainment and amyloid deposition across the spectrum from normal cognition to dementia: neuroimaging evidence for protection and compensation. Neurobiol Aging. 59:72–79.
Arenaza-Urquijo EM, Landeau B, La Joie R, et al. 2013. Relationships between years of education and gray matter volume, metabolism and functional connectivity in healthy elders. NeuroImage. 83:450–457.
Associação Brasileira de Empresas de Pesquisa [Brazilian Association of Research Companies] 2003. Critério de Classificação Econômica Brasil
[Criteria for Economic Classification in Brazil]. http://www.abep.org
. Accessed December 1, 2007.
Bartrés-Faz D, Solé-Padullés C, Junqué C, et al. 2009. Interactions of cognitive reserve with regional brain anatomy and brain function during a working memory task in healthy elders. Biol Psychol. 80:256–259.
Beaulieu C. 2002. The basis of anisotropic water diffusion in the nervous system: a technical review. NMR Biomed. 15:435–455.
Bennett CM, Wolford GL, Miller MB. 2009. The principled control of false positives in neuroimaging. Soc Cogn Affect Neurosci. 4:417–422.
Boltzmann M, Mohammadi B, Samii A, et al. 2017. Structural changes in functionally illiterate adults after intensive training. Neuroscience. 344:229–242.
Brucki SM, Nitrini R, Caramelli P, et al. 2003. Sugestões para o uso do Mini-Exame do Estado Mental no Brasil [Suggestions for the use of the Mini-Mental State Examination in Brazil]. Arq Neuropsiquiatr. 61:777–781.
Buckner RL. 2004. Memory and executive function in aging and AD: multiple factors that cause decline and reserve factors that compensate. Neuron. 44:195–208.
Cao X, Yao Y, Li T, et al. 2016. The impact of cognitive training on cerebral white matter in community-dwelling elderly: one-year prospective longitudinal diffusion tensor imaging study. Sci Rep. 6:33212.
Caramelli P, Barbosa MT, Sakurai E, et al. 2011. The Pietà study: epidemiological investigation on successful brain aging in Caeté (MG), Brazil. Methods and baseline cohort characteristics. Arq Neuropsiquiatr. 69:579–584.
Carreiras M, Seghier ML, Baquero S, et al. 2009. An anatomical signature for literacy. Nature. 461:983–986.
Castro-Caldas A, Petersson KM, Reis A, et al. 1998. The illiterate brain: learning to read and write during childhood influences the functional organization of the adult brain. Brain. 121:1053–1063.
Catani M, Thiebaut, de Schotten M. 2012. Atlas of Human Brain Connections. New York, New York: Oxford University Press.
Ciccarelli O, Werring DJ, Wheeler-Kingshott CA, et al. 2001. Investigation of MS normal-appearing brain using diffusion tensor MRI with clinical correlations. Neurology. 56:926–933.
Dehaene S, Cohen L, Morais J, et al. 2015. Illiterate to literate: behavioural and cerebral changes induced by reading acquisition. Nat Rev Neurosci. 16:234–244.
Dehaene S, Pegado F, Braga LW, et al. 2010. How learning to read changes the cortical networks for vision and language. Science. 330:1359–1364.
Ebly EM, Hogan DB, Parhad IM. 1995. Cognitive impairment in the nondemented elderly: results from the Canadian Study of Health and Aging. Arch Neurol. 52:612–619.
Farfel JM, Nitrini R, Suemoto CK, et al. 2013. Very low levels of education and cognitive reserve: a clinicopathologic study. Neurology. 81:650–657.
Folstein MF, Folstein SE, McHugh P. 1975. ‘Mini-Mental State’: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 12:189–198.
Fontenelle LF, Bramati IE, Moll J, et al. 2011. White matter changes in OCD revealed by diffusion tensor imaging. CNS Spectr. 16:101–109.
Foubert-Samier A, Catheline G, Amieva H, et al. 2012. Education, occupation, leisure activities, and brain reserve: a population-based study. Neurobiol Aging. 33:423; e415–425.
Fratiglioni L, Wang HX. 2007. Brain reserve hypothesis in dementia. J Alzheimers Dis. 12:11–22.
Graham J, Rockwood K, Beattie B, et al. 1997. Prevalence and severity of cognitive impairment with and without dementia in an elderly population. Lancet. 349:1793–1796.
Hagmann P, Jonasson L, Maeder P, et al. 2006. Understanding diffusion MR imaging techniques: from scalar diffusion-weighted imaging to diffusion tensor imaging and beyond. Radiographics. 26(suppl 1):S205–S223.
Hoenig MC, Bischof GN, Hammes J, et al. 2017. Tau pathology and cognitive reserve in Alzheimer’s disease. Neurobiol Aging. 57:1–7.
Hua K, Zhang J, Wakana S, et al. 2008. Tract probability maps in stereotaxic spaces: analyses of white matter anatomy and tract-specific quantification. Neuroimage. 39:336–347.
Jiang H, van Zijl PC, Kim J, et al. 2006. DtiStudio: resource program for diffusion tensor computation and fiber bundle tracking. Comput Methods Prog Biomed. 81:106–116.
Johansen-Berg H. 2010. Behavioural relevance of variation in white matter microstructure. Curr Opin Neurol. 23:351–358.
Jones DK, Knösche TR, Turner R. 2013. White matter integrity, fiber count, and other fallacies: the do’s and don’ts of diffusion MRI. Neuroimage. 73:239–254.
Kanai R, Rees G. 2011. The structural basis of inter-individual differences in human behaviour and cognition. Nat Rev Neurosci. 12:231–242.
Kempermann G, Gast D, Gage FH. 2002. Neuroplasticity in old age: sustained fivefold induction of hippocampal neurogenesis by long-term environmental enrichment. Ann Neurol. 52:135–143.
Kurumaji A, Itasaka M, Uezato A, et al. 2017. A distinctive abnormality of diffusion tensor imaging parameters in the fornix of patients with bipolar II disorder. Psychiatry Res. 266:66–72.
Lavdas I, Miquel ME, McRobbie DW, et al. 2014. Comparison between diffusion-weighted MRI (DW-MRI) at 1.5 and 3 tesla: a phantom study. J Magn Reson Imaging. 40:682–690.
Lecours AR, Mehler J, Parente MA, et al. 1988. Illiteracy and brain damage. 3: A contribution to the study of speech and language disorders in illiterates with unilateral brain damage (initial testing). Neuropsychologia. 26:575–589.
Mesulam M. 2005. Imaging connectivity in the human cerebral cortex: the next frontier? Ann Neurol. 57:6–7.
Nichols T, Holmes A. 2002. Nonparametric permutation tests for functional neuroimaging: a primer with examples. Hum Brain Mapp. 15:1–25.
Nitrini R, Bottino CM, Albala C, et al. 2009. Prevalence of dementia in Latin America: a collaborative study of population-based cohorts. Int Psychogeriatr. 21:622–630.
Nitrini R, Caramelli P, Herrera E Jr, et al. 2004. Performance of illiterate and literate nondemented elderly subjects in two tests of long-term memory. J Int Neuropsychol Soc. 10:634–638.
Nitrini R, Caramelli P, Herrera E Jr, et al. 2005. Performance in Luria’s fist-edge-palm test according to educational level. Cogn Behav Neurol. 18:211–214.
Norton S, Matthews FE, Barnes DE, et al. 2014. Potential for primary prevention of Alzheimer’s disease: an analysis of population-based data. Lancet Neurol. 13:788–794.
Opdebeeck C, Martyr A, Clare L. 2016. Cognitive reserve and cognitive function in healthy older people: a meta-analysis. Neuropsychol Dev Cogn B Aging Neuropsychol Cogn. 23:40–60.
Pajevic S, Pierpaoli C. 1999. Color schemes to represent the orientation of anisotropic tissues from diffusion tensor data: application to white matter fiber tract mapping in the human brain. Magn Reson Med. 42:526–540.
Park S, Ryu SH, Yoo Y, et al. 2018. Neural predictors of cognitive improvement by multi-strategic memory training based on metamemory in older adults with subjective memory complaints. Sci Rep. 8:1095.
Petersson KM, Silva C, Castro-Caldas A, et al. 2007. Literacy: a cultural influence on functional left–right differences in the inferior parietal cortex. Eur J Neurosci. 26:791–799.
Pliatsikas C, Moschopoulou E, Saddy JD. 2015. The effects of bilingualism on the white matter structure of the brain. Proc Natl Acad Sci U S A. 112:1334–1337.
Prince M, Bryce R, Albanese E, et al. 2013. The global prevalence of dementia: a systematic review and metaanalysis. Alzheimers Dement. 9:63–75.
Resende EPF, Tovar-Moll FF, Ferreira FM, et al. 2017. Integrity of white matter structure is related to episodic memory performance in the low-educated elderly. Arq Neuropsiquiatr. 75:778–784.
Rueckert D, Sonoda LI, Hayes C, et al. 1999. Nonrigid registration using free-form deformations: application to breast MR images. IEEE Trans Med Imaging. 18:712–721.
Scarmeas N, Zarahn E, Anderson KE, et al. 2003. Cognitive reserve modulates functional brain responses during memory tasks: a PET study in healthy young and elderly subjects. Neuroimage. 19:1215–1227.
Skeide MA, Kumar U, Mishra RK, et al. 2017. Learning to read alters cortico-subcortical cross-talk in the visual system of illiterates. Sci Adv. 3:e1602612.
Smith SM, Jenkinson M, Johansen-Berg H, et al. 2006. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage. 31:1487–1505.
Smith SM, Jenkinson M, Woolrich MW, et al. 2004. Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage. 23 (suppl 1):208–219.
Smith SM, Nichols TE. 2009. Threshold-free cluster enhancement: addressing problems of smoothing, threshold dependence and localisation in cluster inference. Neuroimage. 44:83–98.
Stern Y. 2012. Cognitive reserve in ageing and Alzheimer’s disease. Lancet Neurology. 11:1006–1012.
Thiebaut de Schotten M, Cohen L, Amemiya E, et al. 2014. Learning to read improves the structure of the arcuate fasciculus. Cerebral Cortex. 24:989–995.
Tovar-Moll F, de Oliveira-Souza R, Bramati IE, et al. 2014. White matter tract damage in the behavioral variant of frontotemporal and corticobasal dementia syndromes. PLoS One. 9:e102656.
United Nations Educational, Scientific and Cultural Organization Institute for Statistics. 2012. International Standard Classification of Education ISCED 2011. Montreal, Quebec, Canada: UIS.
United Nations Educational, Scientific and Cultural Organization Institute for Statistics. 2016. 50thAnniversary of International Literacy Day: Literacy rates are on the rise but millions remain illiterate
. UIS Fact Sheet September 2016, Number 38.
Valenzuela MJ, Sachdev P. 2006. Brain reserve and dementia: a systematic review. Psychol Med. 36:441–454.
Vaqué-Alcázar L, Sala-Llonch R, Valls-Pedret C, et al. 2017. Differential age-related gray and white matter impact mediates educational influence on elders’ cognition. Brain Imaging Behav. 11:318–332.
Wahlund LO, Barkhof F, Fazekas F, et al. 2001. A new rating scale for age-related white matter changes applicable to MRI and CT. Stroke. 32:1318–1322.
Woods RP, Grafton ST, Holmes CJ, et al. 1998. Automated image registration: I. General methods and intrasubject, intramodality validation. J Comput Assist Tomogr. 1:139–152.