To develop sensitive measures of visual memory for the diagnosis of AD, we used electroencephalography (EEG) with entirely novel paradigms, relying on the exact synchronization of the human brain to a visual stimulus repeated at a periodic, fixed rate. This approach – the Fast Periodic Visual Stimulation (FPVS) – provides an objective measure of a process by sampling this process periodically, without requiring an explicit task for the participant, and with an extremely high sensitivity compared to conventional approaches.
We applied the FPVS approach to identify early deficits in face identity discrimination and memory for faces in aged individuals and MCI and AD patients.
Face stimuli were displayed at a constant rate of 6 Hz (F) through sinusoidal contrast modulation from 0% to 100% (Figure 1A). In a first sequence assessing face identity discrimination, one face identity (A) was presented continuously during a 60 s sequence. Every fifth stimulus (i.e F/5 = 1.2 Hz), different oddball face identities were presented, resulting in a sequence of AAAABAAAAC... (Figure 1B). If participants discriminated face A from face B or C, a significant response should appear at 1.2 Hz and multiple integer of this frequency (i.e. harmonics). The following sequence assessed face encoding/consolidation processes by testing face recognition. The facial identity (A) presented during the previous sequence was alternated with the presentation of a novel face identity (i.e. resulting in a sequence of AZAZA) (Figure 1C). The repetition frequency for one face identity was therefore 3 Hz (i.e. F/2). If participants recognized face A, a significant response is expected at 3Hz.
The validation of these experiments in young adults showed that FPVS is suitable to assess face discrimination and face recognition processes. These results were published in an international peer-reviewed journal (Lithfous & Rossion, 2018) and were presented during the Annual Meeting of the International Neuropsychology Society in 2017.
Investigation of face identity discrimination processes in healthy elderly showed that, although significant responses were observed at frequencies of interest, amplitude of the response was significantly reduced compared to young adults. These results suggest slight age-related effects on neural processes involved in face discrimination processes. These results were presented at the Annual Meeting of the Society for Neuroscience in 2016. Face recognition was found to be preserved during normal aging.
Face identity discrimination processes were preserved in AD and MCI, whereas face recognition was impaired. More particularly, the 3Hz response was observed only in 15% of the patients, compared to 80% of the healthy elderly (Figure 2). These results support deficits in implicit memory in AD pathology. Moreover, deficits in implicit memory were observed in mild stage of AD and in MCI, which can be considered as the prodromal stage of the disease, suggesting that deficits in implicit memory appear early in the course of the disease.
