Zhang measured Chinese speech intelligibility under combinations of reverberation and interfering noise, and compared older listeners with younger ones. Under the same acoustic conditions, older people scored 30 per cent lower on average. Age, reverberation time and signal-to-noise ratio all had a significant influence, and the signal-to-noise ratio had the largest of the three. Zhang varied reverberation and signal-to-noise ratio and not any standard. The abstract available here does not report how many people were tested.
Tani followed 49,511 participants of the Japan Gerontological Evaluation Study for three years and took dementia incidence from the public long-term care insurance system, so the outcome came from records rather than from a report. Access to shops selling fruit and vegetables was measured two ways: objectively, by counting stores within 500 metres and one kilometre of home, and subjectively, by asking. Against the best subjective availability, the worst gave a hazard ratio of 1.74 with a confidence interval of 1.49 to 2.04, and the middle categories 1.47 and 1.22, with a significant trend. The objective measure gave 1.60 for the second-lowest quartile. Both remained significant after adjustment for demographics, health status and other neighbourhood features.
Segen had young and older adults navigate a square virtual environment while remembering object locations, with fMRI recording how the brain encoded the distance to the walls. Both age groups remembered objects near a boundary better, and that effect was larger in the older group. The averages matched. The spread did not: reliance on boundaries varied about two and a half times more widely among the older adults, at a standard deviation of 0.005 against 0.002, and Levene’s test confirmed the difference in variance. The mean reliance did not differ between the groups at all.
Huang had 30 young and 30 older adults navigate five mazes, each built so that only one decision strategy would work at the junctions: serial order, associative cue, beacon, relative location or cognitive map. The mazes ran in counterbalanced order across two sessions to limit fatigue. Older participants were less accurate overall, as expected. Contrary to the existing literature, the decline was no steeper for the cognitive-map strategy than for the serial-order or associative-cue ones.
Naveilhan recorded brain activity from electrodes over the scene-selective regions while 28 young and 28 older adults reoriented themselves using objects as landmarks. Older adults performed worse and showed longer latencies in the early markers of visual processing. Making the landmarks smaller impaired both age groups, and delayed the P1 component only in the older one. The variable that produced the delay was the size of the landmark.
Xu randomised 49 older adults to ten minutes of active virtual reality training in a digital copy of a real building, ten minutes of passive video, or nothing, then set them ten wayfinding tasks in the real building. On the tasks they had practised there was no advantage. On new tasks the virtual reality group finished faster, at a hazard ratio of 1.71, and travelled less distance, at 2.03.
Tani followed 49,511 people and took the outcome from insurance records rather than from a questionnaire. Huang tested 60, Naveilhan 56 and Xu 49. Segen scanned 30 older adults against a young comparison group whose size the abstract does not give. Zhang’s listening tests report no sample size in the abstract.