Bilingualism provides a neural reserve for aging populations
Introduction
Cognitive decline is characterized by a slow and progressive decline of memory and cognitive abilities resulting from pathological conditions that cause brain cell death in the aging population. Although treatments are still lacking, studies show that the onset of cognitive decline and dementia can be significantly delayed by intellectual and life-style factors, including education, occupation level and leisure activities. These individual factors contribute to ‘cognitive reserve’ in seniors. Higher levels of cognitive reserve are associated with a reduced risk of developing cognitive decline and a lower rate of memory decline in normal aging (Stern et al., 2006). Stern et al. (2009) suggested that life experiences contribute to individual differences at the brain level, i.e., providing a neural basis for the concept of cognitive reserve (to which we may refer to as ‘neural reserve’). The benefit of intellectual and life-style cognitive reserve is potentially considerable, given the socio-economic and affective impact of cognitive decline in typical aging. Delaying the onset of cognitive decline can lead to substantial economic benefits (Bialystok et al., 2004).
The concept of cognitive reserve is based on multiple observations that particularly intelligent people with a high level of education seem to fare better in the face of cognitive decline, maintaining cognitive function for a longer period of time compared to people with lower levels of education. Epidemiological studies (Garibotto et al., 2008; for a review see Valenzuela and Sachdev, 2006) have demonstrated that low levels of education are strongly correlated with a higher risk of cognitive decline. Aging-induced decline in cognitive abilities is usually associated with structural and functional changes in the brain. For instance, progressive loss of gray matter volume in anterior regions is typically observed (Jernigan et al., 2001, Raz et al., 2005) as is progressive disruption of white matter between anterior and posterior cortical regions (Pfefferbaum et al., 2005, Head et al., 2004, Madden et al., 2009, Gunning-Dixon et al., 2009).
It has been postulated that bilingualism may act as a cognitive reserve. This notion is based upon evidence that executive functions may be better ‘trained’ in bilinguals as compared to monolinguals (Bialystok et al., 2004, Bialystok et al., 2005, Bialystok et al., 2006). Indeed, there is a body of evidence showing that the constant juggling of two (or more) languages in multi- and bi-lingual speakers tunes their conflict resolution capacity, which in turn, transfers to non-linguistic behavioral and cognitive functions. Specifically, domain-general executive functions might benefit from brain plasticity induced by bilingualism, which then allows greater resistance to cognitive decline (Abutalebi et al., 2012, Zou et al., 2012). Brain areas underlying executive control, such as the pre-SMA/ACC (pre-supplementary motor area/ anterior cingulate cortex) and the DLPFC (dorsolateral prefrontal cortex) regions are more highly stimulated in bilingual speakers and this may result in greater cognitive reserve that compensates for the brain atrophy found in normal aging. Indeed, behavioral studies in elderly bilinguals support this notion (Bialystok et al., 2007, Bialystok et al., 2009). For instance, Bialystok et al. (2004) investigated the effect of aging on executive processes in a relatively large group of monolingual and bilingual speakers. Participants were divided into five different age group ranges (30–39; 40–49; 50–59; 60–69; and 70–79) and all performed an attentional task (i.e., Simon task). The study found that bilinguals were faster in resolving the cognitive conflicts on this specific task. Moreover, the advantage for bilinguals was more evident with increasing age. Both monolingual and bilingual speakers show effects of slowing on task performance with age. However, the difference between the two groups widened with age. Several behavioral studies report a ‘bilingual advantage’ on executive functions in almost every age group (infants, children and adults) (Bialystok and Craik, 2010, Kovács and Mehler, 2009). It should be noted that this evidence is not universally accepted (see Paap and Greenberg, 2013) especially for studies focusing on children and young adults. Following children and young adults engage in many cognitively challenging activities that may be at least equivalent to the cognitive challenges provided by bilingualism. However, old adults tend to have fewer cognitively enriching experiences than younger adults, and hence, any putative advantage provided by bilingualism could be more prominent (Valian, 2015). A further hypothesis is that eventual cognitive advantages increase with experience during aging, suggesting age dependent development of cognitive reserve (Hilchey and Klein, 2011).
As to potential neural repercussions, it is well documented that bilingualism induces beneficial experience-related structural changes in terms of increased gray and white matter densities when compared to monolingual speakers. Bilingualism induces structural changes in several brain areas including the left inferior parietal lobule (Mechelli et al., 2004, Della Rosa et al., 2013), the anterior cingulate cortex (ACC) (Abutalebi et al., 2012), and in subcortical structures such as the left caudate (Zou et al., 2012), and putamen (Abutalebi et al., 2013a). These areas are part of the executive control network and this may explain why bilinguals usually have a cognitive advantage in executive control tasks. Interestingly and specifically related to aging, Luk et al. (2011) report that aging bilinguals have globally increased white matter when compared to their monolingual peers. Likewise, also showed that aging bilinguals have increased gray matter densities over the temporal poles and orbito-frontal cortex. These neural data fit relatively well with the recent evidence that bilingualism delays the onset of dementia (Bialystok et al., 2012). For instance, bilingual participants are diagnosed with dementia of the Alzheimer's Type (DAT) about 4–5 years later compared to monolinguals (Schweizer et al., 2012, Craik et al., 2010). These findings were recently confirmed and generalized to other types of dementia (such as vascular dementia) in a large study conducted in India (Alladi et al., 2013).
The main aim of the present study is to investigate the neural basis of the bilingual advantage in a group of aging bilinguals as compared to a matched monolingual control group. For this purpose, subjects carried out an attentional control task (i.e., the Flanker task) and their performance-measured in reaction times (RTs) – was correlated to brain structure. The Flanker task captures cognitive conflict resolution involving attentional control and inhibition processes (Fan et al., 2002). Horizontal black lines with arrowheads pointing to left or right were presented (see Fig. 1a for examples) and participants were instructed to indicate the direction of the central target arrow by pressing a dedicated response button as fast as possible. The central target is presented with congruent, incongruent or neutral flankers. The conflict effect is calculated by subtracting response times to congruent trials from those to incongruent trials (Fan et al., 2002). The Flanker task is widely used in studies of the bilingual cognitive advantage (Abutalebi et al., 2012, Calabria et al., 2011, Costa et al., 2008).
The distribution of RTs in attention tasks is usually positively skewed. However classical measures of central tendency (Gaussian component) neglect this variability and therefore could potentially incur the risk of masking information about the real trend of the dispersion and unique characteristics of the populations sampled, possibly leading to a Type 2 error. To provide a better fit of the RT distribution, the ex-Gaussian function involves a mathematical convolution of a normal (Gaussian) and an exponential distribution, producing the following 3 parameters: μ (mu), reflecting the mean of the Gaussian component of the distribution; σ (sigma), the standard deviation of the Gaussian distribution, and τ (tau), the mean and the standard deviation of the exponential component. The sum of μ and τ yields the total mean of the distribution, whereas adding the square of the standard deviation of σ and τ draws its variance. The leading edge of the distribution is reflected in μ, whereas the skewness of the distribution is better reflected in τ.
Application of the ex-Gaussian function has been successfully employed to compare differences between groups, such as healthy elderly subjects when compared with individuals who have dementia (Spieler et al., 1996, Verhaeghen and Hoyer, 2007). Recently, Tse et al. (2010) showed that healthy aging had clear effects on both μ and τ in a set of attentional control tasks, whereas early stage DAT only had an additional effect on τ. Balota et al. (2010) demonstrated that the slow tail of the RT distribution in the Stroop task may also be useful in predicting divergence from a cognitive normal state to early stage DAT across a 12 year longitudinal follow up study. Jackson et al. (2012) showed a significant association between measures of intra-individual variability (IIV) reflected in the RTs of an attentional control task, the parameter τ, reflecting an exaggeration of the slow tail of the RTs, and white matter volume in an aging population and in early-stage DAT individuals, concluding that IIV and τ are sensitive to breakdowns in executive control processes in healthy and pathological aging. In short, applying the ex-Gaussian function to reaction times allows the splitting of a normal distribution into sub-components (μ and τ) that appear to be independently useful in predicting cognitive decline. It has been argued that between-group differences in the normal component μ are more likely to reflect differences in more “automatic” processes across the whole task, whereas between-group differences in the exponential component τ have been related to more “controlled” processes across the task (such as inhibition) (for a critical review see: Matzke and Wagenmakers, 2009).
Of note, and related to bilingualism, Calabria et al. (2011) used ex-Gaussian analysis to investigate potential cognitive advantages in a large group of young bilinguals given the Flanker task. They found that bilinguals perform faster on both components μ and τ, for incongruent and congruent trials and also had a smaller conflict effect. Calabria et al. also performed a correlation analysis between the magnitude of μ and τ and found that μ and τ were positively correlated for monolinguals whereas in the bilingual group they were not, suggesting that these parameters are independent in bilinguals but not in monolinguals. They argued that the ex-Gaussian analysis was informative for investigating the bilingual advantage in young adults. No study has yet investigated the bilingual advantage in elderly bilinguals using ex-Gaussian analysis.
As aforementioned, in the present study we employed the ex-Gaussian analysis approach in order to analyze RTs measured from the Flanker task in aging bilinguals. Participants also underwent structural MRI scanning to obtain gray matter volume indexes to permit correlations of these results with behavioral performance. Following the results from Calabria et al. (2011), we predicted that bilinguals would perform faster on both components μ and τ, for incongruent and congruent trials and would also show a smaller conflict effect. Furthermore, we predicted differential correlations between components and specific neural structures for cognitive processing in the bilingual and monolingual groups.
Section snippets
Subjects
Thirty healthy bilinguals from Hong Kong (16 Cantonese-English, 14 Cantonese-Mandarin; 13 males; mean age 63.2; standard deviation [SD]±5.86; age range 55–75) and thirty healthy Italian monolinguals (14 males; mean age 61.85; standard deviation [SD]±6.71; age range 49–75) participated in this study. The monolingual control group was chosen from the aging population in Milan because of the relative difficulty of finding real monolinguals in Hong Kong. Usually, the few monolinguals in Hong Kong
Experimental procedures
All participants (bilinguals and monolinguals) performed the Flanker task (Fan et al., 2002). The task was presented on a computer screen. Event presentation consisted of a fixation point appearing at the center of the screen for 400 ms, followed by a row of five horizontal black lines with arrowheads pointing to left or right for 1700 ms (see Fig. 1). Participants were instructed to indicate the direction of the central target arrow by pressing a dedicated response button as fast as possible.
ANOVA on RTs for bilinguals and monolinguals
A 2×2 ANOVA was carried out on RT data using group (bilingual–monolingual) as between subjects factor and congruency (cong versus incong) as within subjects factor. The ANOVA showed a main effect of group [F(1,36)=5.851, p=0.021; =0.14] indicating that bilinguals (mean=670 ms; st dev=20.55) were faster than monolinguals (mean=740 ms; st dev=20.55), and also a main effect of congruency [F (1,36)=284.29, p<0.001; =0.89] indicating that for both groups responses on congruent trials (mean=649
Data acquisition: bilinguals
High-resolution T1 structural images were acquired for the 30 bilinguals at the 3 T MRI center of the University of Hong Kong using a 3 T Achieva Philips MR scanner (Philips Medical Systems, Best, the Netherlands). An axial high-resolution structural MRI scan was acquired for each subject (magnetization prepared rapid gradient echo, 150 slice T1-weighted image, repetition time=8.03 ms, echo time=4.1 ms; flip angle=8°, field of view=250×250, matrix=256, acquisition time (TA)=9.35 min, mode=3D
Structural neuroimaging results
A significant aging effect was found in the right DLPFC (coordinates: x=34.5; y=34.5; and z=16.5) for the whole sample (bilinguals and monolinguals). Namely, a negative correlation was found between age and gray matter volume indicating that an increase in age correlates with a decrease in gray matter volume in right DLPFC. Plots of the parameter estimates for the effects of the two covariates for the μ-incong and τ-cong showed a significant inverse correlation between GM densities in the
Discussion
The aim of our study was to investigate whether bilingualism effectively provides a neural and cognitive reserve for the aging population. For this purpose, we investigated the performance of aging bilinguals and monolinguals on the Flanker task providing a measurement of their attentional control abilities. For both, congruent and incongruent, trials bilinguals outperformed monolinguals. We then performed an ex-Gaussian analysis on the resulting RTs, which provides the advantage of not loosing
References (66)
- et al.
Bilingual language production: The neuro cognition of language representation and control
J. Neurolinguistics
(2007) - et al.
The role of the left putamen in multilingual language production
Brain Lang.
(2013) - et al.
Language proficiency modulates the engagement of cognitive control areas in multilinguals
Cortex
(2013) - et al.
Bilingualism protects anterior temporal lobe integrity in aging
Neurobio. Aging
(2014) - et al.
Effect of bilingualism on cognitive control in the Simon task: evidence from MEG
NeuroImage
(2005) - et al.
Bilingualism as a protection against the onset of symptoms of dementia
Neuropsychologia
(2007) - et al.
Bilingualism: consequences for mind and brain
Trends Cogn. Sci.
(2012) - et al.
Conflict monitoring and anterior cingulate cortex: an update
Trends. Cogn. Sci.
(2004) - et al.
Bilingualism aids conflict resolution: evidence from the ANT task
Cognition
(2008) - et al.
A neural interactive location for multilingual talent
Cortex
(2013)