Snowdon et al. (1996). Linguistic Ability in Early Life and Cognitive Function and Alzheimer’s Disease in Late Life: Findings from the Nun Study.
Arhe Vaninetti
Research Methods and Statistics, PSY150, Pacific University, Oregon
March 1, 2023
Snowdon et al. (1996) aimed to find if linguistic ability was associated with Alzheimer’s disease or declined cognitive function in later life. In this study, the researchers convenience sampled from the School Sisters of Notre Dame congregation, and found ample participants, but narrowed it down to 93 nuns who had similar demographic characteristics and had written autobiographies in their early life. Commonly, after a nun had spent 4 years in training, they would often write an autobiography before they took their vows—which Snowdon et al. then used to characterize which nuns had high/low linguistic ability through linguistic complexity and idea density.
The scores for linguistic ability were calculated using the last 10 sentences of the autobiographies, and with the Developmental Level metric. The two individuals who did the scoring for linguistic ability were blinded to the participant’s writing samples to ensure consistency. For these coders, the interrater reliability was 0.88 for idea density, and 0.93 for grammatical complexity. Of the 47 Sisters who lived to see the year 1995, the researchers had them write new autobiographical statements, which they then compared with the original statements from their youth. For this comparison, the linguistic scores were highly correlated (0.73 for idea density and 0.62 for grammatical complexity).
Next, the researchers gave the 47 living Sisters a standard neurological function battery, which tested various cognitive functions—and thus tested for Alzheimer’s disease. For the Sisters that were deceased, the researchers counted neurofibrillary plaques and tangles in the hippocampus, Broddman areas, part of the parietal lobe, and the middle frontal lobe. The prevalence of the plaques and tangles were calculated using standard neurological fields.
Of the 93 Sisters, 34 had low idea density, and 32 had low grammatical complexity (e.g., a ‘low’ score being in the bottom 1/3rd of the sample). Regarding the 14 Sisters that had died during the study, 5 had Alzheimer’s disease when the brain matter was viewed under a microscope. The researchers go on to say that of the 5 Alheimer’s positive Sisters, 100% of them had low idea density, but because the sample was so small and had little power, the researchers looked at an additional 5 Sisters’ brain matter, which showed signs of Alzheimer’s disease, from a similar convent. These individuals had low idea density in 80% of the cases.
When comparing the Alzheimer’s stricken brain matter with normal brain matter, the Alzheimer’s brains had low idea density 90% of the time, and the neurologically healthy brains had low idea density 13% of the time (p = <0.001). There was a weak association between low grammatical complexity and Alzheimer’s disease.
The study concluded with the results that low idea density in early life is associated with higher rates of Alzheimer’s disease in later life. This study provided strong subjects that lacked confounding factors, as the Sisters lived similar lifestyles and had structured days. This lack of confounding variables adds to the power of the study, and was an ingenious sample to select.
An interesting dimension that could be added to this study would be the use of modern neurological instruments, like a PET scan, which would make the results easier to obtain and produce a larger sample size. Furthermore, the Sisters would not have to donate their brains to science, and it would streamline the funerary process for the families.
Since estrogen is seen as a neuroprotectant, it would add to the results if the Sister’s hormone levels could be monitored and/or tested regularly throughout the study. It is possible that estrogen could have produced a confound or moderator in the relationship between idea density and Alzheimer’s disease.