Frontal lobe epilepsy: an eye tracking study of memory and attention

Objective To explore the characteristics and mechanisms of working memory impairment in patients with frontal lobe epilepsy (FLE) through a memory game paradigm combined with eye tracking technology. Method We included 44 patients with FLE and 50 healthy controls (HC). All participants completed a series of neuropsychological scale assessments and a short-term memory game on an automated computer-based memory evaluation platform with an eye tracker. Results Memory scale scores of FLE patients including digit span (U = 747.50, p = 0.007), visual recognition (U = 766.50, p = 0.010), and logical memory (U = 544.00, p < 0.001) were significantly lower than HC. The patients with FLE took longer to complete the four levels of difficulty of the short-term memory game than healthy controls (level 1: U = 2974.50, p = 0.000; level 2: U = 3060.50, p = 0.000; level 3: U = 2465.00, p = 0.000; level 4: U = 2199.00, p = 0.000). During the memory decoding period, first fixation on the targets took significantly longer for FLE patients for all difficulty levels compared to controls (level 1: U = 3407.00, p = 0.008; level 2: U = 3618.00, p = 0.036; level 3: U = 3345.00, p = 0.006; level 4: U = 2781.00, p = 0.000). The average fixation duration per target among patients with FLE was found to be significantly longer compared to HC (level 1: U = 2994.50, p = 0.000; level 2: U = 3101.00, p = 0.000; level 3: U = 2559.50, p = 0.000; level 4: U = 2184.50, p = 0.000). The total fixation duration on AOI/total completion time of FLE patients was significantly lower than those of HC for levels 1 to 3 (level 1: U = 1557.00, p = 0.000; level 2: U = 2333.00, p = 0.000; level 3: U = 2757.00, p = 0.000). Furthermore, the eye tracking data during the memory decoding phase were correlated with neuropsychological scale scores (p < 0.05). Conclusion Patients with FLE exhibited short-term memory impairment probably due to deficits in attentional maintenance, especially during the memory decoding phase. Eye tracking technology provided the possibility to help separate and quantify visual attention from memory processing, contributing to exploring underlying mechanisms of memory impairment in FLE.


Introduction
Frontal lobe epilepsy (FLE) is the second most prevalent focal epilepsy in adults, accounting for approximately 20-30% of cases (Giovagnoli et al., 2020;McGonigal, 2022).It is increasingly recognized as a brain network disorder, which impacts a broad range of cognitive domains (Rayner et al., 2015;Nair and Szaflarski, 2020).Although attention deficits and impaired executive functioning are relatively common, memory deficits have also been commonly described among patients with FLE (Kibby et al., 2019).However, the frontal lobe is not traditionally the brain area responsible for memory function (Buschman and Miller, 2007;Centeno et al., 2010;Al-Aidroos et al., 2012;Fiebelkorn and Kastner, 2020).Therefore, how memory can be impaired in FLE and whether patients have specific memory characteristics remain inconclusive.
Previous studies have demonstrated that attention plays a crucial role in the process of memory formation (Zelinsky and Loschky, 2005;Chun and Johnson, 2011;Bahmani et al., 2019).More specifically, visual attention fundamentally determines which information is inputted and encoded into memory (Hollingworth et al., 2001;Voss et al., 2017).Rock and Gutman (1981) have proven that people can better recall items that have received more visual attention.Bahmani et al. (2019) also confirmed the inextricable relationship between memory and visual attention at the electrophysiological, behavioral, and anatomical levels.However, the conventional scales of memory evaluations, which include the Wechsler Memory Scale (WMS), the Hopkins Verbal Learning Test, and the Rey Auditory Verbal Learning Test, are unable to separate visual attention from memory and lack the ability to gauge and accurately measure it (Sherman et al., 2012;McAuley et al., 2015;Sveikata et al., 2019).Consequently, a tool that can quantitatively measure visual attention during memory processes and assist us in investigating the influence of attention deficit on memory impairment is urgently needed.
Eye tracking technology has the ability to capture visual behaviors and track the path of an individual's eye movements in real time as patients engage in cognitive tasks (Boraston and Blakemore, 2007;Oyama et al., 2019;Tao et al., 2020).When integrated with memory tasks, it can accurately and comprehensively quantify visual attention (Boraston and Blakemore, 2007;Blais et al., 2008).Sylvia B. Guillory applied eye tracking-based measurement in the study of memory function among individuals with Autism Spectrum Disorder (ASD), specifically those with Phelan-McDermid Syndrome (PMS) (Guillory et al., 2021).They revealed that patients with PMS and co-morbid ASD diagnoses have poorer memory and lower attentional engagement with social images (Guillory et al., 2021).Using oculomotor tracking combined with a visual paired-comparison paradigm, patients with Rett syndrome showed restricted and immature visual attention characteristics during recognition memory tasks (Rose et al., 2013).In our previous work on an automated computer-based memory assessment platform with an eye tracker, we found that TLE patients experienced memory retrieval difficulties with relative sparing of attention (Zhu et al., 2021;Xiao et al., 2023).
In the present study, we hypothesize that patients with FLE might have a specific memory impairment profile, which is different from those of TLE, and attention might contribute to deficits in the memory process in FLE.To this end, we first applied the digit span, visual recognition, and logical memory scales, as per the Wechsler Memory Scales-Chinese Revision (WMS-RC), in patients with FLE.To further delineate the patterns of memory impairment, we employed an automated memory assessment platform, which, relying on eye tracking technology, recorded and analyzed associated parameters in both memory encoding and decoding phases, including total completion time, first fixation time on target, average fixation duration per target, etc.We made a correlation analysis between the eye tracking index and the performance of memory scales.Our work showed that patients with FLE exhibited impaired short-term memory function, which might be mostly due to deficits in attention maintenance, especially during the memory decoding phase.

Participants
A total of 44 FLE patients [22 male,age 28.11 (9.15)] and 50 healthy controls (HC) [22 male, age 30.22 (12.39)] were enrolled in our study and were matched for age, sex (male/female), and years of education.All FLE patients were recruited from September 2020 to August 2021 at Xiangya Hospital, Central South University, and were diagnosed with FLE according to the definition of the International League Against Epilepsy (ILAE) by two trained epileptologists (Fisher et al., 2017).The diagnostic criteria for FLE were based on history (typical symptoms of seizures suggestive of frontal lobe origin), EEG, and neuroimaging.Exclusion criteria: (1) those aged under 16 years or over 60 years; (2) those who had a history of neurological disorders other than epilepsy and a history of neurosurgery; (3) those who claimed to have subjective memory impairment; (4) those who also had psychiatric disorders or comorbidities; (5) those who failed the eye tracking calibration procedure or were unable to understand the experimental procedure.

Memory scale assessment
All the participants undertook a memory scale assessment.We applied digit span, visual recognition, and logical memory scales, as per the Wechsler Memory Scales-Chinese Revision (WMS-RC), to assess the memory ability of the participants.
Digit span aims to measure verbal working memory and can be conducted forward and backward.In digit span forward, the participants are required to repeat a string of digits in sequence after hearing them, starting by repeating 3 digits, with a maximum of 11 digits, and there are two chances for each digit length.In digit span backward, the participants are required to repeat the digits in reverse, starting with 2 digits, with a maximum of 10.The test was terminated when 11 digits were correctly recalled or incorrectly recalled twice in a row.
Visual recognition measures visuospatial memory.In this task, participants were shown eight cards for 30 s and were required to recall the eight cards in 28 cards.The cards included graphics, Chinese characters, and mathematical symbols.The more cards the participant recognized, the higher the score.
Logical memory task measured verbal episodic memory.The participants were asked to browse a short story, and the investigator would read it aloud at the same time.When the investigator finished reading, the story card would be taken back and the participants were required to immediately recall the story in as much detail as possible.There were two different short stories for each trial.The more details the participant recalled, the higher the score.

Memory paradigm
An automated computer-based memory assessment platform was adapted from Li et al., as reported in our previous study, for measuring short-term memory.The stimuli included 38 front-facing human images downloaded from the Chinese University of Hong Kong (CUHK) student database 1 and 38 images of fractals downloaded from the Web.
The participants were presented with the image(s) for 5 s (referred to as the encoding phase); then, the image(s) would disappear and 12 images in 3 rows and 4 columns would be presented on the screen (referred to as the decoding phase).They were required to choose the image they memorized by clicking the mouse, they could not proceed with the next trial until all the correct targets had been clicked.There are four difficulty levels in the short-term memory game, starting with memorizing one target (level 1) and ending with memorizing four memory targets (level 4), with four trials in each level (Figure 1).The terms "YES" and "NO" are presented as feedback for the correct selections and incorrect selections, respectively.

Eye tracking task 2.4.1 Eye tracking recording
Concurrently with the assessment of short-term memory, we gathered real-time eye tracking information utilizing the EyeLink 1,000 Plus eye tracker to observe the visual search conducted by the 1 http://mmlab.ie.cuhk.edu.hk/archive/facesketch.html participants throughout the task.The eye tracking data was captured in monocular remote mode with a sampling rate of 500 Hz.To ensure accuracy, participants were instructed to minimize movement of their upper body while using the remote mode.Minor movements of the head and upper body were permitted, as the oculomotor system can be automatically calibrated.A standard 9-point calibration procedure was implemented to guarantee the precision and accuracy of the collected data.Participants who were unable to pass the calibration process were excluded from the subsequent experiment.

Eye tracking analysis
Eye tracking data were analyzed with DataViewer (version 4.2).We extracted the eye tracking indicators for the encoding and decoding phases separately.Fixation, saccade, and other behaviors were directly calculated using the EyeLink system.The EyeLink system uses three thresholds to detect saccades and fixation: motion, velocity, and acceleration, corresponding to 0.15 degrees, 30 degrees/s, and 8,000 degrees/s 2 , respectively.The update interval and data accumulation period of the whole system are set at 50 milliseconds.The eye tracking indicators included total completion time, first fixation time on target, average fixation duration per target, average fixation count per target, total fixation duration on area of interest (AOI)/total completion time, and average fixation duration per target/total fixation duration on AOI.First fixation time on target refers to the time point when the participants notice the presented target.The reaction time refers to the duration between the introduction of a stimulus and the subsequent response.The first fixation time on target in the encoding phase can be used to reflect the subject's response and alertness when participating in the experiment.Average fixation duration per target is the average fixation durations within per target.Average fixation count per target is the average number of fixations that fall within the target.Total fixation duration on AOI/total completion time is the percentage of total fixation duration on AOI to total completion time.Average fixation duration per target/total fixation duration on AOI is the percentage of average fixation duration per target to the total fixation During the trial, a target was shown on the screen for 5 s (encoding), followed by the appearance of 12 (4 × 3) potential answers.Participants were required to recall and click on the correct target from the 12 objects (decoding).The targets were numbered from 1 to 4, and participants were required to choose all the correct targets during the decoding phase in order to proceed to the next level.The orange circle indicates the target and the white circle in the decoding phase is the interference graphic.
10. 3389/fnins.2023.1298468Frontiers in Neuroscience 04 frontiersin.orgduration on AOI.We used Python (version 3.7.10) to process the eye tracking data.For each participant, the eye tracking indicators were averaged over the trials to obtain the average indicator corresponding to each difficulty level.We also extracted the above eye tracking indicators under memorized fractal images and memorized frontfacing images.

Statistical analysis
The statistical analysis was conducted using the SPSS software package (version 27.0; SPSS Inc., Chicago, Illinois, United States).In order to analyze categorical variables such as sex in both FLE patients and healthy controls, χ2 analysis was employed.Continuous variables such as age and years of education, which passed the normality test, were analyzed using an independent samples t-test.The data were presented as mean and standard deviation (Mean ± SD).For memory scale scores and eye tracking data that significantly deviated from normal distribution, the Mann-Whitney U test was used to compare the differences between the two groups.Spearman's correlation analysis was used to determine the correlation between the eye tracking indicators and the memory scale score.Two-tailed p-values were calculated for all tests, and p-values less than 0.05 were considered statistically significant.

Participant demographics
A total of 44 adult FLE patients and 50 HC eligible for this study were included.The demographic and clinical characteristics of the participants are shown in Table 1.No statistically significant differences were found between FLE patients and HC regarding sex (p = 0.561), age (p = 0.341), and years of education (p = 0.079).

Memory scale assessment
As shown in Table 2, there were significant differences between FLE patients and HC in the scores of scales reflecting memory function, including digit span (U = 747.50,p = 0.007), visual recognition (U = 766.50,p = 0.010), and logical memory (U = 544.00,p < 0.001).The FLE patients' scores in these scales were significantly lower than those of HC.

Short-term memory task
There was a significant difference in total completion time between the two groups in the decoding phase in the four difficulty levels (level 1: U = 2974.50,p = 0.000; level 2: U = 3060.50,p = 0.000; level 3: U = 2465.00,p = 0.000; level 4: U = 2199.00,p = 0.000).The total completion time of the FLE patients in the four difficulty levels was significantly longer than that of the HC (Figure 2A).

Indicators of the encoding phase
In the memory encoding phase, the first fixation time on targets of FLE patients was significantly longer than that of HC in level 4 (U = 3603.5,p = 0.032), while no significant difference was found in the remaining difficulty levels.There were no differences in average fixation duration per target and average fixation count per target in levels 1 to 4 (all p > 0.05).When it comes to different image stimuli, no differences were found in all eye tracking indicators in the four difficulty levels under both fractal image stimuli and front-facing image stimuli (Table 3).

Indicators of the decoding phase
The first fixation time on targets of FLE patients was significantly longer than that of HC in the four difficulty levels (level 1: U = 3407.00,p = 0.008; level 2: U = 3618.00,p = 0.036; level 3: U = 3345.00,p = 0.006; level 4: U = 2781.00,p = 0.000).The average fixation duration per target of FLE patients was significantly longer than that of HC (level 1: U = 2994.50,p = 0.000; level 2: U = 3101.00,p = 0.000; level 3: U = 2559.50,p = 0.000; level 4: U = 2184.50,p = 0.000), the average fixation count per target of FLE patients were also significantly higher than that of HC (level 1: U = 3206.50,p = 0.001; level 2: U = 3320.00,The total fixation duration on AOI/total completion time of FLE patients was significantly lower than that of HC in levels 1 to 3 (level 1: U = 1557.00,p = 0.000; level 2: U = 2333.00,p = 0.000; level 3: U = 2757.00,p = 0.000), while no difference was found in level 4 (level 4: U = 3683.00,p = 0.054).The average fixation duration per target/ total fixation duration on AOI of FLE patients was significantly higher than that of HC in level 1 (U = 3987.00,p = 0.020), while no differences were found in the remaining difficulty levels (Table 4 and Figure 3).

Correlation of memory scale assessment and eye tracking indicators
The total completion time of the decoding phase was negatively correlated with the scores of the digit span (rs = −0.525,p = 0.000), visual recognition (rs = −0.315,p = 0.002), and logical memory (rs = −0.398,p = 0.000) (Figure 6).

Discussion
In this study, we confirmed the presence of memory deficits among FLE patients in combination with scales and short-term memory tests by eye tracking technology.We found that FLE patients showed intact attentional alertness but impaired attentional retaining in memory processes.In addition, these differences in eye tracking metrics appeared mainly during the memory decoding phase and performed worse with both fractal and front-facing image stimuli.Poor performance in memory scales associated with the eye tracking data suggests that eye tracking technology has the potential to be used as a supplementary neuropsychological tool for memory evaluation.
It has been widely accepted that patients with FLE do possess memory impairments, showing dysfunction during encoding, free recall, and retrieval (Exner et al., 2002, Nolan et al., 2004), and there is also evidence of memory deficit following frontal lobe resection for epilepsy (McDonald et al., 2001).In our study, FLE patients performed worse on both the WMS-RC test and the eye tracking-based shortterm memory tests, confirming the view that memory deficit is one of the long-term effects of FLE on the neural networks involved in memory function.The medial prefrontal cortex, along with an intact connection with the medial temporal lobe and hippocampus, has been involved in memory encoding and retrieval (Centeno et al., 2010;Squire et al., 2015;van den Berg et al., 2021).Petrides and Pandya (1988) have identified a larger group of fiber projections in the frontal cortex that originated from the various areas of the temporal region, which might be the anatomic basis of the memory circuit.The rich interconnectivity between the temporal and frontal lobes may facilitate epileptic activity propagation and subsequent dysfunction in distant structures.Additionally, a series of functional neuroimaging findings have revealed that the frontal lobe participated in supporting effective memory in conjunction with memory-associated structures like the hippocampus (Centeno et al., 2012;Braakman et al., 2013).
Therefore, we inferred that the memory deficits in FLE might be partly due to the dysfunction of different regions within the frontal lobes contributing to long-term memory functioning (Patrikelis et al., 2009;Kibby et al., 2019;Caciagli et al., 2023).The frontal lobe is the hub of the attentional cognitive control network, also involved in the orchestrating functions of attention.Thus, whether the memory impairment among patients with FLE was dependent on attention deficit was further investigated.We found no difference in performance between FLE and HC in the first fixation time on target in the memory encoding phase, which means the attentional alerting is mostly retained in FLE to guarantee that they are in a heightened sensitive state and have the ability to respond fast  and correctly (Posner and Petersen, 1990).However, the following sustained attention during the retrieval period has been identified as obviously impaired due to a lower ratio of the total fixation duration on AOI/total completion time among patients with FLE.Previous studies have shown that the frontal cortex not only produces regulatory signals related to attention but also functions to maintain and control spatial attention (Fiebelkorn and Kastner, 2020).Thus, we inferred that the recurrent epileptic discharges and pathology in different frontal regions would interfere with the attentional network and further negatively affect higher cognitive functions such as attention and memory (Rayner et al., 2015;Fiebelkorn and Kastner, 2020).We also noticed that the patients with FLE showed impairments primarily in recognition memory, which aligns with previous reports by Centeno et al. (2010).They observed that poor memory performance in FLE was caused by difficulties in information retrieving, rather than encoding dysfunction (Centeno et al., 2010), resulting in repeatedly searching for the correct target, corresponding to longer fixation duration and more fixation counts on the target captured.Why these difficulties in recalling memories mostly contributed to memory deficits in FLE patients is of interest.We believe a possible reason is that the damaged prefrontal cortex from FLE disturbs the memory decoding process.The hippocampus is responsible for the initial memory encoding, while the prefrontal cortex takes charge of memory retrieval and consolidation (Preston and Eichenbaum, 2013;Hainmueller and Bartos, 2020).During the memory recall phase, the prefrontal cortex receives visual input, integrates prior knowledge and short-term memory from the hippocampus, and ultimately regulates attentional allocation (Buschman and Miller, 2007;Katsuki and Constantinidis, 2014).Consequently, when the frontal lobe and affective cognitive network exhibit structural or functional abnormalities, memory recollection in the decoding phase suffers.Accumulating evidence has supported this hypothesis by showing that patients with right frontal lobe tumors performed poorly in the retrieval process (Anderson et al., 2011).Our study involved six patients diagnosed with FLE, specifically experiencing frontal lobe absence seizures.Drawing from previous research on absence seizures, we hypothesize that abnormal discharges originating in the frontal lobe contribute to cognitive impairment and disturbances in consciousness among these patients (Lenkov et al., 2013;Wang et al., 2023).In forthcoming research, the integration of eye movement techniques with advanced methods like functional magnetic resonance imaging (fMRI) and magnetoencephalography (MEG) could further explore the influence of the frontal lobe on absence seizures.
There is also a significant correlation between the total completion time of the memory game and the scores of the scales.Previous studies have found that the eye tracking index correlates well with cognitive performance.Therefore, we believe that eye tracking technology combined with cognitive tasks could be applied as an alternative option for neuropsychological assessment in clinical practice and long-term monitoring.Considering our finding of short-term memory deficit, probably due to attentional maintenance dysfunction during memory retrieval, future psychological training strategies, which focus on cognitive adjusting to sustain attentional focus, are suggested to cope with those problems and to improve memory.Nevertheless, there are several limitations in our current study.Firstly, it was a clinical study conducted at a single center with small sample size; prospective studies with larger participation with longterm follow-up are required.Secondly, lateralization of FLE would have an effect on cognitive performance.For better screening of the memory deficit profile of FLE, future studies should take lateralization into consideration.Moreover, the impact of anti-seizure medications on cognition is an inevitable confounding factor that limits our statistical power, newly diagnosed patients should be enrolled to exclude the interfering effects of drugs.Additionally, we can design new memory game paradigms that combine fractal pictures and frontal-facing images, which can facilitate further investigation of the connection and difference between general and social cognition.

Conclusion
Our study innovatively combined traditional scales and eye tracking short-term memory tasks, to explore the characteristics and mechanisms of working memory impairment in patients with FLE.We found that short-term memory deficits in FLE patients are probably due to attentional maintenance dysfunction during the memory retrieval phase.Eye tracking technology with cognitive tasks could be a reliable potential supplementary neuropsychological tool for cognition evaluation and precise process intervention.

FIGURE 2
FIGURE 2Comparison of (A) total completion time of the Short-Term Memory task, (B) total completion time under fractal image stimuli, and (C) total completion time under front-facing image stimuli between the HC group and FLE group at four difficulty levels.Statistical significance for each task between groups is indicated by asterisk(s) (*p < 0.05, **p < 0.01, ***p < 0.001).HC, healthy control; FLE, frontal lobe epilepsy.

FIGURE 3
FIGURE 3Comparison of (A) first fixation time on target, (B) average fixation duration per target, (C) average fixation count per target, and (D) total fixation duration on AOI/total completion time of the decoding phase in all tasks between the HC group and FLE group.Statistical significance for each task between groups is indicated by asterisk(s) (*p < 0.05, **p < 0.01, ***p < 0.001).HC, healthy control; FLE, frontal lobe epilepsy.

FIGURE 5
FIGURE 5Comparison of (A) average fixation duration per target, (B) average fixation count per target, and (C) total fixation duration on AOI/total completion time of the decoding phase in the front-facing image stimuli task between the HC group and FLE group.Statistical significance for each task between groups is indicated by asterisk(s) (*p < 0.05, **p < 0.01, ***p < 0.001).HC, healthy control; FLE, frontal lobe epilepsy.

FIGURE 4
FIGURE 4Comparison of (A) first fixation time on target, (B) average fixation duration per target, (C) average fixation count per target, and (D) total fixation duration on AOI/total completion time of the decoding phase in the fractal image stimuli task between the HC group and FLE group.Statistical significance for each task between groups is indicated by asterisk(s) (*p < 0.05, **p < 0.01, ***p < 0.001).HC, healthy control; FLE, frontal lobe epilepsy.

FIGURE 6
FIGURE 6Correlations between the scores of memory scale assessment (digit span, visual recognition, logical memory) and TCT, FFT on AOI, FFT on target, AFD per target, AFC per target, TFD on AOI/TCT, AFD per target/TFD on AOI.Statistical significance for each task between groups is indicated by asterisk(s) (*p < 0.05, **p < 0.01, ***p < 0.001).TCT, total completion time; FFT, first fixation time; AOI, area of interest; AFD, average fixation duration; AFC, average fixation count; TFD, total fixation duration.

TABLE 1
Demographic and clinical characteristics of participants.

TABLE 2
Memory scale assessment.

TABLE 3
Comparison of eye tracking indicators in encoding of participants.FLE (n = 44) HC (n = 50) test p-value FLE, frontal lobe epilepsy; HC, healthy control.

TABLE 4
Comparison of memory game and eye tracking indicators in decoding of participants.