Rationale:
Face-name association (FNAME) is an episodic memory task that can detect age- related memory decline, cortical amyloid burden, and mild cognitive impairment (MCI). We adapted the FNAME to test immediate and 30-minute delayed recall in temporal lobe epilepsy patients (TLE) and healthy controls (HCs). We hypothesized that TLE patients, who often have hippocampal dysfunction, would have poorer FNAME performance than controls. We collected eye-tracking data to determine if encoding and retrieval processes differed between TLE and HCs
Methods:
We recruited TLE patients and HCs from NYU Comprehensive Epilepsy Center from 2018-2025. HCs were eligible if (1) Aged 18-60, (2) MOCA≥ 27, and (3) fluent in English. Patients were eligible if (1) MOCA≥ 24 and (2) Had a probable or definite diagnosis of TLE, determined by a combination of EEG, MRI Brain, and/or seizure semiology. Subjects participated in FNAME, which showed 12 pairs of color faces and written name. After a brief distraction task, subjects were shown the 12 faces in randomized order, then asked to recall their associated names and after a 30- minute delay.
Results:
Sixty-three (63) subjects were enrolled, including 24 HCs, 39 TLE patients (26 LTLE, 13 RTLE). HCs and TLE patients were matched in age (p = 0.008) , gender (p = 0.153), handedness (p=0.198), education (0.969) and Montreal Cognitive Assessment (MOCA) scores (p = 0.224). TLE patients, compared to HCs, remembered fewer face-name associations in the immediate (TLE = 41.96 % vs. Healthy = 59.72% faces remembered, p = 0.0019) and 30-minute delay condition (TLE = 37.8 % vs. Healthy =58.33%, p = 0.0013). Lower MOCA scores, specifically among LTLE patients (F(1,16) = 15.55, p = 0.0003), correlated with poorer face-name associative recall. Nine subjects (4 TLE, 5 HC) subjects underwent eye tracking. A greater number of saccades were observed during correct encoding compared to incorrect encoding (Correct vs Incorrect, p = 0.052). However, correct recall was characterized by fewer saccades compared to incorrect recall in TLE patients (Correct vs. Incorrect p = 0.016). A significant increase in HC pupil dilation was observed after face stimulus presentation of incorrectly recalled trials (t(3) = -3.28, p = 0.046).
Conclusions:
Immediate and delayed 30-minute FNAME recall can discriminate between TLE patients and HCs, and correlated with MOCA score. Correct encoding is characterized by more frequent saccades, whereas correct retrieval is associated with fewer saccades, compared to incorrect trials. Face name association recall performance can detect memory impairment in clinical populations.
Funding:
R01 NS1007806 (AL)
1K23NS104252 (AL)
FACES (AL)
Department of Neurology (AL)