Oxford University Press's
Academic Insights for the Thinking World

Of language, brain health, and global inequities

One of the greatest public health challenges of our century lies in the growth of neurodegenerative disorders. Conditions such as Alzheimer’s disease, Parkinson’s disease, and frontotemporal dementia stand as major contributors to disability and mortality in affluent and under-resourced nations alike. Currently affecting over 55 million individuals, their prevalence is expected increase significantly by 2050—especially in less developed countries, where risk factors are most impactful and mainstream clinical approaches least developed.

Language research in the fight against neurodegeneration

Against this background, researchers from various fields are searching for new, affordable, and scalable digital innovations to facilitate diagnosis and other clinical tasks across the globe. Speech and language assessments have emerged as crucial tools, offering robust insights for detecting, characterizing, and monitoring these diseases. For instance, individuals with Alzheimer’s often struggle with word retrieval, experience difficulties in constructing grammatically complex sentences, and exhibit challenges in understanding or expressing figurative language. These linguistic deficits appear in early and preclinical disease stages, differentiate Alzheimer’s from other forms of dementia, allow predicting the onset of core symptoms, and even capture brain anomalies that typify the disorder.

These clinical applications can be boosted through artificial intelligence tools. New digital technologies allow capturing specific alterations in recorded or written language samples in a non-invasive, patient-friendly, and cost-effective way. Such is the type of solution required to reduce clinical disparities across low-, middle-, and high-income countries. Multicentric research initiatives, large grants from leading funding agencies, and science-based companies are spearheading exciting projects to validate and expand this novel framework. However, a critical challenge looms large: the lack of linguistic diversity in the field threatens its scalability and undermines its potential for more equitable testing worldwide.

Disorders of language vs. disorders of languages

The field is marked with inequities. Less than 0.5% of the world’s 7,000 languages have received any attention in this research field. Also, although English is spoken by roughly 17% of the world’s population, it accounts for nearly 70% of all published studies on speech and language in neurodegeneration. Moreover, large language models and feature extraction tools are available for only a handful of languages. Of course, none of this would be a major issue if links between language anomalies and brain dysfunctions were universal across the world’s languages—if that were the case, we could rely on the abundant findings from English and apply them to patients worldwide, irrespective of their language. Unfortunately, the reality is much more complicated.

As it happens, cross-linguistic differences deeply influence the presentation of speech and language symptoms, challenging the universality of existing diagnostic criteria and candidate disease markers. For instance, a sentence production study showed that Italian-speaking persons with Alzheimer’s could be identified by their tendency to omit subjects, a phenomenon notably absent in their English-speaking counterparts. The distinction lies in the inherent structure of the languages. Unlike English, Italian allows deducing sentence subjects from verb conjugations (the Italian verb ‘camminiamo’ inherently implies a first-person plural subject, whereas the English verb ‘walk’ requires a preceding ‘we’ to convey the same meaning). More strikingly, linguistic anomalies may be diametrically opposed between languages. For example, research on Alzheimer’s shows that different pronouns (words like ‘I’, ‘their’, ‘ours’) tend to be overused among English-speaking patients and underused in Bengali-speaking patients—relative to healthy speakers of the same languages. This, too, likely reflects differences between both languages, as Bengali grammar includes many more (and morphologically more complex) pronouns than English. Succinctly, the linguistic markers that may signal a given disease among speakers of one language may not be relevant among speakers of another language.

Taking action

These findings underscore the need to consider language diversity when examining the linguistic impact of neurodegenerative conditions. Such is the call we raised in our recent article in Brain (García et al., 2023). Researchers must broaden the representation of languages, incorporating diverse linguistic communities to identify shared and distinguishing properties. Multicentric collaborations, harmonized protocols, and cross-linguistic tools must be forged for a more inclusive and comprehensive understanding of neurodegeneration across regions and cultures. The path forward requires overcoming core challenges, such as establishing robust pipelines for comparing outcomes across languages, disentangling linguistic and non-linguistic sources of heterogeneity, and securing funds for language research across underrepresented regions. Ideally, local-global connections should be prioritized to integrate country-specific needs and resources with leading worldwide trends.

Promisingly, strategic efforts are being made in this direction. Consider, for example, the International Network for Cross-Linguistic Research on Brain Health (Include). Supported with initial funds from the Global Brain Health Institute, the Alzheimer’s Association, and the Alzheimer’s Society, Include aims to foster trans-regionally equitable approaches to language-based neurodegeneration research. The network has grown continually since its launch in November 2022. It now has over 140 members spanning 80 centers sites across 30 countries. Five network-wide projects are being run, targeting diverse phenomena across multiple languages in large cohorts of persons with Alzheimer’s, Parkinson’s, and frontotemporal dementia variants. Include is also leading awareness-raising actions, such as the Language Diversity and Brain Health webinar series, hosted in collaboration with the Bilingualism, Languages, and Literacy Special Interest Group of the Alzheimer’s Association’s Diversity and Disparities Professional Interest Area. Initiatives like these can make a difference towards fairer language-based research on brain dysfunctions.

The bottom line

Speech and language assessments hold a valuable key to unlocking generalizable insights on neurodegeneration. To harness their full potential, however, we must bridge the linguistic gap in research, embracing more diverse samples and more inclusive practices. These actions are vital to ensure that valuable tools for equitable brain health assessments do not turn into a new source of global inequity.

Feature image by Studioroman via Canva.

Recent Comments

There are currently no comments.