A Loss for Words: Distinct Language Impairments in Alzheimer’s Disease
Through a cross-sectional analysis of the NACC dataset, researchers have discovered that early-onset and late-onset Alzheimer’s disease cause distinctly different language impairments. The work, published in Alzheimer’s and Dementia, clears up a longstanding puzzle in the field and points toward clinical interventions that could improve the quality of life for people with Alzheimer's disease and caregivers.
Language impairments, including trouble recalling words, an inability to form coherent sentences, and general difficulty communicating, are common features of dementia. Unfortunately, it’s a hard problem to study. Because language requires a complex set of skills, researchers and speech pathologists use multiple tools to measure it, but that limits the scope of studies. Administering several complex tests to each participant and comparing the results is expensive and slow.
“In the rehabilitation space, there is a lot of work towards contextualizing what standard assessment outcomes might actually mean for a person, but that typically focuses on really small cohorts just based on feasibility,” says Jeanne Gallée, PhD, CCC-SLP, a researcher and speech-language pathologist in the Department of Medicine at the University of Washington, and lead author on the new paper. She adds that “setting up a larger treatment study or even just a single time point study with individuals with specific diagnoses can be quite challenging.”
An alternative approach is to look at larger cohorts of patients using simpler measurements, but that ignores important distinctions between different types of language impairments. “Language has historically been looked at as a single measure, so we have the black box of language, and that doesn’t really tell us much,” says Gallée. Different kinds of language problems stem from different limitations and call for different interventions, so these broad-brush analyses seldom provide practical insights, regardless of their statistical power.
Indeed, this conflict between detailed studies that are too small and large assessments that lack specificity has yielded confusing results. For example, while some researchers have found significant differences in language impairments between early- and late-onset Alzheimer’s disease populations, others have found no difference between the two groups. “This is where the NACC dataset came in pretty serendipitously, where you just have access to such a rich and [deep] dataset where you can apply some of these questions,” says Gallée. In particular, the NACC data include multiple types of language assessments across thousands of participants.
Leading a team of collaborators from the University of Washington and several other institutions, Gallée was able to perform a cross-sectional analysis of more than 1,000 people with early-onset Alzheimer’s disease and over 4,000 with the late-onset form. The team focused on spoken language generation at the time of Alzheimer’s disease diagnosis. “[Those are] typically just the kinds of assessments that people always perform across study sites, and it is a nice proxy to thinking about day-to-day conversation,” says Gallée. Individual words carry most of the meaning in spoken language, so while assessments of grammar might be interesting, recalling nouns and verbs serves as a good indicator of someone’s ability to communicate.
“NACC data include multiple types of language assessments across thousands of participants”
After researchers had selected their study population and the relevant measurements, the analysis was relatively straightforward. “The NACC dataset is made beautifully available to participating researchers and it’s super well-organized, so it felt like it went very smoothly after making those initial decisions,” says Gallée.
The data revealed that participants with early-onset Alzheimer’s disease had more severe language impairments in several areas, compared to those with late-onset disease. This finding suggests the two forms of Alzheimer’s disease have distinct profiles in language performance, clearing up the earlier confusion in the literature and suggesting a need for new clinical approaches. “That has super important implications for not only diagnosis, but long-term care planning; we know that communication is an essential function of everyday quality of life, participation, autonomy, caregiver burden, but we don’t have the care procedures that are representative of that at the time of diagnosis,” says Gallée.
Detailed language tests such as the ones in the NACC dataset are not practical in routine clinical care, though. To address that, Gallée is now building on her NACC-based findings and other work to develop a new assessment tool, which simplifies the process of testing a person’s language abilities while simultaneously gathering deeper diagnostic information. “The aim is for this to be a tool that can be implemented at routine clinical visits, assessment, monitoring, and ongoing care journeys, where you end up having quantified profiles of communication in nuanced aspects of language, all based on a single conversation,” she says.
Gallée J, Gibbons LE, Choi SE, Lee M, Scollard P, Trittschuh EH, Mez J, Saykin AJ, Foldi NS, Mukherjee S, Crane PK. Facets of language performance in early-onset and late-onset Alzheimer's disease dementia. Alzheimers Dement. 2025 Sep;21(9):e70705. doi: 10.1002/alz.70705. PubMed PMID: 40990522; PubMed Central PMCID: PMC12458901.