Memory characteristics in mesial temporal lobe epilepsy: Insights from an eye tracking memory game and neuropsychological assessments

Abstract Aims To compare different patterns of memory impairment in patients with two subtypes of mesial temporal lobe epilepsy (MTLE) and healthy controls. Methods Thirty‐five healthy controls and 41 patients with MTLE were recruited, of which 25 patients were diagnosed as hippocampal sclerosis (HS‐MTLE), and the rest 16 patients were lesion‐negative (MRI‐neg MTLE). Participants completed the Wechsler memory assessment and a short‐term memory game on an automated computer‐based memory assessment platform with an eye tracker. Results Both the MRI‐neg MTLE and HS‐MTLE groups took longer time to complete the short‐term memory game than healthy controls (p < 0.001, Cohen's d = 1.087; p = 0.047, Cohen's d = 0.787). During the memory encoding phase, the MRI‐neg MTLE group spent significantly shorter time than healthy controls on the difficult levels with three (p = 0.004, Cohen's d = 0.993) and four targets (p = 0.016, Cohen's d = 0.858). During the memory decoding phase, the HS‐MTLE group spent less time looking on the targets compared to controls when recalling and finding four targets (p = 0.004, Cohen's d = −0.793), while the MRI‐neg MTLE group spent significantly longer time on the distractors and shorter time on the region of interests (ROIs) for all difficulty levels (all p < 0.05) than controls. Furthermore, the eye tracking data were correlated with the scores of the Wechsler Memory Scale after Bonferroni correction (p < 0.05). Conclusion Patients with MRI‐neg MTLE demonstrate impaired memory mostly due to attention deficits, while those with HS‐MTLE show memory impairment with relative sparing of attention. Eye tracking technology has the potential of facilitating the investigation of the mechanism of memory defect in MTLE and can serve as a supplementary neuropsychological tool for clinical diagnosis and long‐term monitoring.


| INTRODUC TI ON
Mesial temporal lobe epilepsy (MTLE), the most common type of drug-resistant epilepsy in adults, 1 usually originates from medial temporal lobe and hippocampal regions. 2 Patients with MTLE usually present with various cognitive deficits with memory impairment. 3,4 Although progressive impairment of memory reportedly contributes to difficulties in daily routines in patients with MTLE, 5 researches focusing on how memory encoding and decoding are affected are lacking.
Functional magnetic resonance imaging (fMRI) studies based on memory tasks have shown that the activation of hippocampus is significantly enhanced during memory encoding activities. 6,7 The hippocampus holds an essential role in memory formation and consolidation, and it forms a reciprocal signal circuit crucial for translation of temporary hippocampal output to permanent memory. 8 Hippocampal sclerosis and atrophy are considered responsible for the inability to form new memories, anterograde amnesia, and specific memory deficits. 9,10 Mesial temporal sclerosis (MTS), also commonly referred to as hippocampal sclerosis (HS), is the most common pathological change of MTLE in adulthood. 11 Memory impairment in MTLE is thus probably due to structural abnormalities. 12,13 An alternative view is that patients with MTLE are prone to develop memory impairment regardless of lesion status. 14,15 A recent study suggests that MTLE patients without any lesions on MRI (MRI-neg MTLE) are also associated with memory impairment similar to MTLE patients with HS (HS-MTLE). 15 An fMRI study conducted by Vaughan et al. found that MRI-neg MTLE and HS-MTLE had distinct patterns of network abnormalities. 16 These results support the notion that memory impairment in MTLE may not always be associated with structural abnormalities but rather an intrinsic deficit of the underlying network malfunction.
To date, the characteristics and underlying mechanism of memory impairment in MTLE remain elusive. Although cognitive scales have been widely applied in memory evaluation in individuals, they still have limitations in early diagnosis and long-term monitor due to its high degree of subjectivity, low sensitivity, and reproducibility. 17 In addition, traditional memory scales fail to separate the attention domain from memory formation, while the effect of visual attention on the results of memory tests cannot be ignored, giving rise to mixed and ambiguous conclusions. 18 Eye tracking technology is featured by its millisecond-level sampling rate, quantitative measurements, and reliability during the visual monitoring process and has been widely used in cognitive studies. 17,19 Moreover, this technology can help separate visual attention from memory processing, 20,21 unfold the pattern of memory encoding and decoding, and elucidate the underlying mechanism of memory deficits in diseases. 22,23 In the present study, we used the eye tracking-based automated memory assessment platform, 24 in combination with Wechsler Memory Scale-Revised of China

| Wechsler memory assessment
To assess the memory function of the participants, we applied the Digit Span task, the Visual Recognition task and the Logical memory task of the WMS-RC. 26,27 Digit Span task: This task measured the verbal working memory and, comprises digit span forward and backward subtests. In the Digit Span Forward subtest, after hearing a sequence of digits dictated by the experimenter, the participant was instructed to repeat the sequence of digits. The test ended until either the maximum length was completed or two consecutive incorrectly repeated sequences of same length occurred. In the Digit Span Backward subtest, the participant was asked to repeat the sequence in reverse order.
Visual Recognition task: This task measured the visual-spatial memory. In this task, eight cards were presented to the participants for 30 s, the participants then were asked to recall the eight presented cards in all 28 cards. These cards include graphics, Chinese characters, and mathematical symbols. The more cards the participants identify, the higher the score.  Note: Data presented as n or mean ± SD. The scores of the Digit Span were calculated by summing the forward and backward subtests and then converting them to equivalence scale scores. The scores of the Logical memory were obtained by converting the sum of scores into equivalent scale scores. The scores of the Visual recognition were summed from the correct recognition of the cards and then converted into equivalent scale scores. Statistical significance between groups is indicated by bold styling and asterisk(s) (*p < 0.05, **p < 0.01, ***p < 0.001).
Abbreviations: HS-MTLE, mesial temporal lobe epilepsy with hippocampal sclerosis; MRI-neg MTLE, MRI-negative mesial temporal lobe epilepsy; y, years; SD, standard deviation; Digit Span, the score of the Wechsler memory scale-Digit Span; Logical memory, the score of the Wechsler memory scale-Logical memory; Visual recognition, the score of the Wechsler memory scale-Visual recognition.
the participant was asked to recall as many details as possible immediately. The more details the participants recalled, the higher the score.

| Game design
The automated computer-based memory assessment platform was adapted from Li et al. as reported in our previous study, 24  four columns on the screen (Figures 1, 2). The decoding phase has a time limit of 46 s, after which the game moves on to the next trial.
The subjects were asked to acknowledge the memory target(s) by clicking a mouse. They were prompted for a response and after their response was recorded, feedback was provided and they clicked the left mouse button to continue to the next trial. The time taken to complete each level of the memory game and the number of incorrect trials were recorded, and the average completion time was calculated.

| Eye tracking
Eye tracking data were collected at a sampling rate of 300 Hz and analyzed with Tobii Pro Lab (version 1.123). The region of interests F I G U R E 1 Short-term memory game with Eye Tracking. After calibration, one target is presented for 6 s on the screen (Encoding), and then, the target disappears and 12 (4 × 3) potential answers instead. The participant needs to recall and click on the target from 4 × 3 objects (Decoding). Memory target(s) from one to four in turn, and the participant must select the target(s) from the previous screen in order to proceed to the next level.
time on ROIs during the memory decoding phase divided by the game completion time.

| Demographics
Demographic and clinical characteristics of participants were summarized in Table 1. There were no significant differences in gender, age, or education level among the three groups (p > 0.05). The two MTLE groups did not differ in age at epilepsy onset, duration of epilepsy, seizure frequency of epilepsy, or number of antiepileptic drugs (p > 0.05). and Figure 3).  Table 1 and Figure 3).

| Wechsler memory assessment
We did not find a significant group difference in the performance of the Visual Recognition task among the three groups (H (2) = 5.023, p = 0.081) ( Table 1 and Figure 3).

| Short-term memory game and eye tracking
Through automated computer-based assessment with eye tracking, were not significant (all p > 0.05) ( Table 2 and Figure 4C).  Figure 4D).
During decoding, eye tracking data revealed that the MRI-neg MTLE group spent more time on distractors than controls at all four difficulty levels (all p < 0.05,  Figure 4E).
The percentage of total visit time on the ROIs was significant different among the three groups for all four difficulty levels (all p < 0.05, Table 2). The percentage of total visit time on the ROIs of the MRI-neg MTLE group was significantly less than that the healthy control group for all difficulty levels (all p < 0.05, Table 2). There were no differences between HS-MTLE group and MRI-neg MTLE group in the percentage of total visit time on the ROIs when recalling one, two, or three targets (all p > 0.05), but the difference when recalling and finding four targets is significant (p = 0.012, Cohen's d = +1.323) ( Table 2 and Figure 4F). The differences between the HS-MTLE group with the healthy controls were not significant in the percentage of total visit time on the ROIs for all four difficulty levels ( Table 2 and Figure 4F).

| Correlation of WMS-RC and eye tracking
We

| DISCUSS ION
This is the first study to investigate characteristics of memory impairment in MTLE subtypes using eye tracking technology. We confirmed that MTLE patients exhibited memory impairment based   We propose the following possible mechanism for the different characteristics of memory deficits in the two MTLE groups.

HS-MTLE (n = 16)
Attention usually interfered with the performance of memory tasks.
Attention requires normal neocortical activation via direct and indirect ascending excitatory projects from subcortical regions. 36  We propose that eye tracking-based cognitive tests are a promising supplementary tool for neuropsychological evaluation, especially in the early diagnosis of and quantitative assessment for the cognitive function.

| LI M ITATI O N S
Our present work is limited in several ways. Firstly, the results of this study are limited to the sample size of the study. Prospective studies with larger sample sizes are required to further our understand-

| CON CLUS ION
In conclusion, by the eye tracking-based memory tasks and traditional assessment scales, we find that patients with MRI-neg MTLE are characterized by impaired memory with significantly attention deficit, while the patients with HS-MTLE demonstrate isolated memory impairment. Our results also suggest that eye tracking technology can be a supplementary clinical tool to provide insight into cognitive processes, and holds promise for early diagnosis and intervention in cognitive impairment associated with neurological diseases.

ACK N OWLED G M ENTS
Foremost, we thank all patients and study participants.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflict of interest.