Interesting. It's a finding that needs to be replicated. Although the sample size was fairly large (317) the recovery rates were lower than what we should expect to see in the general population (>75%), but the recovery rates they observed were 55% for monolingual speakers and 25% for bilingual speakers. Those look like big numbers, but I'm skeptical that the difference in recovery rates are due to sampling bias (e.g., I'm not sure we should make grand statements based on a single study on 69 bilingual speakers. Certainly something to look into more, however!)