University of Toronto researchers led by Tom Schweizer have found that Alzheimer’s disease (AD) has to progress twice as far in bilingual patients as in monolingual people before symptoms become apparent. In other words, multilingual people can sustain much more brain damage before they become noticeably symptomatic.
The researchers ran CT scans of matched patients who had been diagnosed with probable AD. Half the patients were fluent in two languages, the other half spoke only one language. Although both groups performed equally well in all cognitive tests, the bilingual subjects displayed twice as much brain atrophy as the monolingual subjects. This data corroborates previous observations that bilingual patients develop AD up to five years later than their monolingual cohorts.
It’s not yet established whether there’s something specific about language or if it’s simply the act of using our brains (and alternating smoothly between two languages in both thought and speech does require an enormous amount of computing power) that is so protective. If speaking two languages is critical, it may not be necessary to have spoken both languages from early childhood.
To be clear, whatever mechanism is at work merely postpones the diagnosis of AD, it doesn't cure or prevent the disease. Still, five extra years of being one's self is nothing to scoff at.