Plasticity of visual evoked potentials in patients with neurofibromatosis type 1

Objective: The inability to properly process visual information has been frequently associated with neurofibromatosis type 1 (NF1). Based on animal studies, the cause of cognitive disabilities in NF1 is hypothesized to arise from decreased synaptic plasticity. Visual cortical plasticity in humans can be investigated by studying visual evoked potentials (VEPs) in response to visual stimulation. Methods: VEP plasticity was assessed by measuring the increase of the peak amplitudes C1, P1, and N1 induced by 10-min modulation of checkerboard reversals in 22 adult NF1 patients and 30 controls. VEP signals were recorded pre-modulation, during modulation, and at 2, 7, 12, 17, 22, 27 min post-modulation. Results: The P1 amplitude increased significantly comparing post-modulation to pre-modulation in the control group. This potentiation was not observed in the NF1 group. Conclusions: Visual cortical plasticity could be measured using VEPs in response to visual stimulation in the control group. Individuals with NF1 may have reduced visual cortical plasticity, as indicated by their non-potentiated response to VEP induction. These findings should be interpreted with caution due to high inter-subject variability. Significance: The present study contributes to an improved assessment of the feasibility for using neurophysiological outcome measures in intervention studies of cognitive deficits among patients with NF1.


Introduction
Neurofibromatosis type 1 (NF1) is associated with cognitive deficits and learning disabilities that can affect quality of life (Varni et al., 2019). In addition to the somatic symptoms associated with NF1, patients have a lower than average intelligence quotient score, attention deficits, impairments in motor learning, and visual information processing difficulties (Hyman et al., 2005;Descheemaeker et al., 2013). Given the cognitive impairments in the visuospatial and visuoperceptual domains, the inability to properly process visual information might contribute to some of the learning disabilities in NF1 (Descheemaeker et al., 2013;Ribeiro et al., 2015). However, it is unknown whether there is primary dysfunction of visual pathways in NF1 adults, and if there are neurophysiological deficits in the visual cortex that could contribute to the cognitive disabilities in NF1.
The underlying cause of the cognitive disabilities in NF1 might be a result of decreased synaptic plasticity, which was found in animal models of NF1 (Costa et al., 2002;Cui et al., 2008;Shilyansky et al., 2010;Omrani et al., 2015). The neurobiological process leading to enduring enhancement of strength or efficacy of synaptic transmission, i.e. long-term potentiation (LTP), is essential for learning and memory. Previous LTP studies were mostly limited to animal studies or surgically excised human cortical tissue, hampering a translation to clinical studies. Non-invasive neurophysiological methods have filled this gap and can measure changes in cognition and learning (Tinga et al., 2019). These measurements include eventrelated potentials in response to sensory stimulation in the human brain. Ribeiro et al. (Ribeiro et al., 2014) observed that, in response to visual stimuli, event-related potentials were already atypical at baseline for late evoked responses in 12 NF1 children, indicating alterations in highlevel processing of visual stimuli in NF1. Furthermore, they found an increased amplitude of alpha brain oscillations in the visual cortex in NF1 patients. The enhancement of alpha brain oscillations is associated with decreased excitability and may be associated with attention problems in visual processing (Ribeiro et al., 2014).
Perceptual learning involves the plasticity of responses to sensory stimulation in the primary sensory cortices. Specifically, Frenkel et al. (Frenkel et al., 2006) showed visual cortical plasticity of the responses to repeated visual stimulation in awake mice. They measured chronic visual evoked potentials (VEPs), which is a type of event-related potential. The measurements showed a time-dependent increase in VEP amplitude in response to a repeated visual stimulus, which disappeared with the presentation of a novel visual stimulus. Visual cortical plasticity in the human visual cortex can be measured by changes in the amplitude of VEPs (Teyler et al., 2005). In psychiatry, the study of VEPs has been used to improve the understanding of the physiology and pathology of several disorders including depression, schizophrenia and bipolar disorder (Normann et al., 2007;Cavus et al., 2012;Elvsåshagen et al., 2012).
VEPs can be elicited in the visual cortex by visual stimulation, for which a flash of light or pattern reversal of the black and white blocks in a checkerboard pattern is typically used (for an overview of studies see Table 1 of (Valstad et al., 2020)). Teyler et al. (Teyler et al., 2005) were the first to demonstrate an increase of one of the components of the VEP in unaffected controls after repetitive visual stimulation using checkerboard reversals. Prolonged stimulation by exposure to flashing light at a high frequency or a 10-min block of checkerboard reversals at a low frequency has been shown to induce potentiation of the VEPs (Normann et al., 2007;Cavus et al., 2012). Changes in VEP amplitudes seem to be more sensitive to checkerboard reversals (Normann et al., 2007;Cavus et al., 2012;Kloppel et al., 2015;Zak et al., 2018).
Potentiation of VEPs has been observed in healthy adults as indicated by a decrease in amplitude of the prominent negative component at 75ms (C1 or N75), and an increase in amplitude of the positive component at 100 ms (P1 or P100) after stimulation (Kloppel et al., 2015). In contrast, in 40 depressed patients, P1 did not increase after a 10-min modulation block of checkerboard reversals while it did in a group of 70 healthy controls (Normann et al., 2007).
In addition, Zak et al., (Zak et al., 2018) observed a significant increase in the peak-to-peak amplitude of P1 to N1 in controls, but not in patients with bipolar disorder type II. The N1 or N135 is the negative component following P1. Collectively, these studies show that VEP induction with checkerboard reversals can indicate deficits in potentiation in the visual cortex.
Clinically, VEPs are used to assess the function of the visual pathway from the eye to the occipital cortex. In NF1 patients, a few VEP studies have been performed in which VEPs were studied under baseline conditions without the induction of VEP plasticity. These studies showed abnormal VEPs at baseline in 26-51% of NF1 patients, including children (aged 6-16 years), adolescents (aged 10-18 years), and adults (aged 18-56 years), compared to controls (Jabbari et al., 1985;Iannaccone et al., 2002;Ammendola et al., 2006;Yerdelen et al., 2011).
More specifically, NF1 patients exhibited a delayed latency of the P1. Additionally, a recent study showed a decreased amplitude of the P1 in 26 NF1 adults compared to controls (Nebbioso et al., 2020). These findings suggest a primary abnormality of the visual pathways in NF1.
Notably, optic nerve gliomas are very common in NF1 and could have influenced the VEP latencies in the previous studies. However, the number of patients in these studies with optic gliomas was low (5-15%; (Ammendola et al., 2006;Yerdelen et al., 2011)) and patients with optic gliomas were often excluded (Jabbari et al., 1985;Iannaccone et al., 2002;Nebbioso et al., 2020). This indicates that delayed VEP in NF1 cannot be fully explained by the presence of gliomas.
To investigate the plasticity of the visual cortex in NF1 patients by assessing VEP potentiation, we studied VEPs using checkerboard reversals at baseline (i.e. pre-modulation), during 10-min modulation and 30-min post-modulation. VEP plasticity was measured by change in the peak amplitude of the VEP signal compared to pre-modulation. To our knowledge, this is the first study that investigates VEP plasticity in NF1, which might be a novel neurophysiological outcome measure associated with cognitive disability.

Subjects
In this study, 22 patients with NF1 and 31 controls between 18-55 years participated after they gave their written informed consent. According to the in-and exclusion criteria, subjects were included if they had no visual problems or neurological illness that involved the visual system. Furthermore, all subjects had no optic nerve gliomas, no MRI (magnetic resonance imaging) lesions, and were without any other ocular pathology based on a general health questionnaire. Subjects had no history or current presence of neurological or psychiatric disorders and did not use psychoactive agents at the time of the study. Patients with NF1 were outpatients from the ENCORE NF1 expertise center for neurodevelopmental disorders at the Erasmus University Medical Center, Rotterdam. Patients with NF1 had a genetic and/or clinical diagnosis of NF1. Controls matched for age and gender were unaffected unrelated peers of the patients or recruited through online advertisements. The Medical Ethics Review Committee of the Erasmus Medical Center Rotterdam approved the study (MEC-2020-0095), which was conducted following the Declaration of Helsinki (2013).

Procedures
VEP recordings took place in the afternoon between 12 PM and 5 PM at the Department of Clinical Neurophysiology at the Erasmus University Medical Center Rotterdam. Subjects were seated in a comfortable chair during the VEP recordings while maintaining focus on a red fixation dot on a screen located 54 cm in front of the subject. We recorded from Oz, Cz, and a reference electrode on the forehead (ground) according to the 10-20 system of electrode placement (Klem et al., 1999). The impedance between electrode and scalp was minimized by injecting conductive, non-alcoholic, viscous gel (OneStep Cleargel, H + H Medizinprodukte GbR, Münster, Germany) in the electrodes. VEP signals were recorded using a NicoletTM Viking EDX system (Natus Neurology Incorporated, Middleton, Wisconsin, USA) with settings according to the ISCEV guidelines (Odom et al., 2016). We used a classical cathode ray tube (CRT) stimulator with a mean photopic luminance of 45 Cd/m2. The mean luminance of the visual stimulator was constant during the checkerboard reversals and identical to the grey screen presented during the intervals. The light in the room was dimmed. We used a custom programmed Raspberry PI to facilitate accurate timing of the protocol. In addition, before the start and at the end of the VEP recording, the level of sleepiness was reported using the Karolinska sleepiness scale (KSS), a self-report questionnaire on a nine-point Likert scale (Akerstedt and Gillberg, 1990).

VEP measurements
VEPs were elicited by checkerboard reversals at a low frequency of 1.92 Hz (Normann et al., 2007;Cavus et al., 2012;Odom et al., 2016). The checkerboard reversal stimuli were presented at a distance of 54 cm with a NICOLET 2015 visual stimulator connected to a ViewSonic E70fB 17-inch display. This setup met the criteria as defined by ISCEV guidelines (Odom et al., 2016). We used a field size of 33 x 25 degrees with large checks of 1x1°. The mean luminance at the center was between 40 and 60 Cd/m 2 and had a Michelson contrast ratio larger than 80%. We accepted variation up to 30% from the center to the periphery. The Viking EDX had a sampling frequency of 48 kHz and the checkerboard pattern changed every halfsecond. In each stimulation block, 40 sweeps were presented within 20 seconds to both eyes.
The responses to the sweeps were averaged. An identical checkerboard reversal was presented continuously for 10 minutes during the modulation block. A grey screen was shown during the intervals between the stimuli. During the experiment, signals were analogous band-pass filtered of 0.05 to 100 Hz and amplified according to the ISCEV guidelines (Odom et al., 2016). Traces (epoched data with one sweep locked to the pattern reversals) exceeding 130 µV were considered blink artifacts and were discarded.
We recorded the mean VEP signals in the stimulation blocks at pre-modulation, modulation, and post-modulation (Normann et al., 2007;Elvsåshagen et al., 2012;Kloppel et al., 2015) (Fig. 1). Pre-modulation, the stimulation blocks with a duration of 20 seconds each, started 1 and 3 min after the start of the experiment (i.e. B1 and B2). We used two measurements at baseline for better data stability. If the two measurements showed no significant difference in amplitude or latency, the average of the two was used in the paired t-tests as pre-modulation in comparison to the average of post-modulation (T1-T6) in accordance with previous literature (Elvsåshagen et al., 2012;Kloppel et al., 2015;Zak et al., 2018). The modulation with a duration of 10 min started 5 min after the start of the experiment. We recorded VEP signals 10 times for 20 seconds each (i.e. M1 to M10) to observe the data stability during modulation.
Post-modulation, the stimulation blocks with a duration of 20 seconds each, started at 2 (T1), 7 (T2), 12 (T3), 17 (T4), 22 (T5), and 27 (T6) minutes after the end of the modulation (Fig. 1). VEP plasticity is measured by a change in the peak amplitudes of the P1, C1, or N1 component when comparing the average of the post-modulation, or the individual time points (T1-T6) of the post-modulation, to pre-modulation.

Statistical analysis
VEP data were analyzed using MATLAB R2019b (Mathworks). A trial was the average of 40 traces with 18 trials over time per individual (i.e. B1, B2, M1-M10, T1-T6). The trials started 50 ms before the onset of each checkerboard reversal and continued for 450 ms after onset (Elvsåshagen et al., 2012). The trial data was baseline corrected (-50 -0 ms prior to stimulus) and digitally low-pass filtered at 48Hz (Heinrich and Bach, 2001;Di Russo et al., 2002;Elvsåshagen et al., 2012). We performed peak detection semi-automatically using a custom-made MATLAB script. We calculated the mean of the VEPs from all recordings per subject to create a subject average VEP signal (Supplementary Figure 1a). In this subject average VEP signal, P1 was identified as the maximum amplitude between 90 and 130 ms, C1 as the last minimum preceding P1, and N1 as the first minimum following P1. Subsequently, peaks were automatically detected from each recorded time point (i.e. trial) by finding the minimum/maximum within a 20ms window surrounding the subject average peak latencies.
Peaks were manually detected if no minimum/maximum was found within this 20ms window.
Amplitudes of the C1, P1 and N1 peaks were calculated to the 50ms baseline. Two experimenters ran the analysis independently and came to a consensus of the exclusion of VEPs or individual peaks that could not be identified (see example in Supplementary Figure 1b). If VEPs or individual peaks could not be identified within a subject, the whole trial was excluded from the analyses.
Statistical analyses were performed using IBM Statistics SPSS (version 25). We tested whether there were differences between groups regarding the confounding variables gender, age, and educational attainment with a Chi-squared test and non-parametrically with Mann-Whitney U test, respectively. Correlations between age, educational attainment (following the ISCED; (Schneider, 2013)), and the main outcomes were evaluated using Pearson or the nonparametric Spearman's correlation coefficients, and p-values were corrected for multiple testing with the Holm-Bonferroni correction. Significant data had a p-value ≤ 0.05.

The VEP response
The peak latencies and amplitudes of the stimulation blocks pre-modulation (i.e. B1 and B2) were tested with a paired t-test. The differences between the two groups in mean peak latencies and amplitudes pre-modulation were tested with independent t-tests.

VEP plasticity
We tested the overall effect of the modulation in the NF1 and control group with a repeated measures ANOVA between-within subject design with factor time (pre-modulation, T1-T6) and factor group. In the ANOVA, we included the first baseline measurement to compare the VEPs acquired using the same number of traces.
Subsequently, a within-subject repeated-measures ANOVA was used to measure the effect of the factor time (pre-modulation, T1-T6) on mean VEP amplitudes in controls and patients separately if a significant interaction effect was established with the overall repeated measures ANOVA (Elvsåshagen et al., 2012). Degrees of freedom were corrected using Greenhouse-Geiser estimates of sphericity.
In addition, we tested the change per peak amplitude of averaged pre-and postmodulation with paired t-tests in controls and patients separately if a significant interaction effect was established with the overall repeated measures ANOVA (Elvsåshagen et al., 2012).

VEP response at M10
We tested the effect of modulation without any delay by comparing the VEP amplitudes of the last minute during continuous visual stimulation (i.e. M10, Fig. 1) with the first baseline measurement using overall repeated measures ANOVA with a between-within subject design.
Subsequently, a within-subject repeated-measures ANOVA was used to measure the effect of the factor time in controls and patients separately if a significant interaction effect was established with the overall repeated measures ANOVA. The data stability during modulation was tested with a repeated measures ANOVA to find any significant differences in VEP amplitude during modulation.

Data availability
Data are available from the corresponding author on reasonable request.

Results
In total, 42 patients with NF1 agreed to eligibility screening, of which 22 subjects were enrolled. Subjects were excluded due to psychiatric comorbidities (n = 7), participation in other clinical trials (n = 2), use of psychoactive medication (n = 3), ocular pathology (n = 2), general health issues (n = 2), epilepsy (n = 2), or cancelled due to COVID-19 (n = 2) (Fig. 2). In addition, 35 eligible subjects without NF1 were screened for study eligibility, of which 31 subjects were enrolled. During the measurements, the number of excluded traces exceeding 130 microV was not significantly different between group (U = 337, p = 0.9, excluded traces: control = 0.21%; NF1 = 0.15%). After the measurements, subjects were excluded if VEPs could not be identified (n control = 1; n NF1 = 2), and trials (18 trials per individual; B1, B2, M1-M10, T1-T6) were excluded if they could not be identified (2.7% of total trials in controls; 3.5% of total trials in patients with NF1). No significant differences were found between patients and controls in age (Mean ± SD: M patient = 29.2 ± 11.4, M control = 33.17 ± 11.2; U age = 161.0, p = 0.10), gender (χ 2 = 0.35, p = 0.56) or level of sleepiness at the start and end of the experiment (U start = 261, p = 0.55, U end =162, p = 0.16). We did find an expected difference in the level of education, for which patients had a significantly lower level of education than controls (U = 127.5, p < 0.001) ( Table 1).

VEP plasticity
The overall repeated measures ANOVA with a between-within subject design showed no overall effects of time (pre-modulation, T1-T6) or group for C1 (F time (6,270) = 1.6, p = 0.2; In the NF1 group, the within-subject repeated measures ANOVA showed no significant effect of the factor time for P1 (F P1 (6, 108)=1.3, p = 0.3) (Fig. 3).
In contrast to the control group, the P1 peak amplitudes were not significantly different between pre-and averaged post-modulation in the NF1 group (t(19)=-0.3, p=0.4).

Discussion
VEP plasticity offers a non-invasive metric to quantify cortical plasticity in the visual cortex. Our findings showed that VEPs were potentiated in control subjects in response to a 10min block of visual stimulation. The P1 amplitude of the VEP between post-modulation and pre-modulation was significantly increased in control subjects. In contrast, this amplitude was not potentiated in response to modulation in the NF1 group, which might suggest deficits in visual cortical plasticity in adults with NF1.
In contrast to our expectations, the latencies and amplitudes of the VEP components pre-modulation were not significantly different between groups. Previous studies of VEP characteristics in patients with NF1 showed a delayed latency and reduced amplitude of the P1 at baseline (Jabbari et al., 1985;Iannaccone et al., 2002;Ammendola et al., 2006;Yerdelen et al., 2011;Nebbioso et al., 2020). In the present study, we did not find abnormalities in VEP components pre-modulation, which suggests a lack of abnormalities of the visual pathways in our NF1 sample. In contrast to previous studies, all subjects were without ocular pathology, optic gliomas or visual problems.

VEP plasticity in the control group
Our findings support the hypothesis that VEP plasticity could be used to identify visual cortical plasticity in humans. The underlying mechanism of VEP plasticity induced by a modulation block has been shown to resemble characteristics of synaptic plasticity. More specifically, VEP plasticity in mice was inhibited by manipulation of N-methyl-D-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors (Frenkel et al., 2006). The activation of these receptors is important for induction of long-term potentiation at cortical synapses (Malinow et al., 2000).
Our findings are in line with previous results of VEP plasticity in humans showing that prolonged modulation by exposure to a 10-min block of checkerboard reversals at a low frequency induces VEP plasticity in unaffected controls as indicated by the changes in peak amplitudes of the VEP (Normann et al., 2007;Elvsåshagen et al., 2012;Kloppel et al., 2015;Zak et al., 2018;Valstad et al., 2020). In these studies, C1 amplitude was significantly decreased, or trended towards significant, in the control group and P1 amplitude increased between post-modulation and pre-modulation. Notably, however, Elvsåshagen et al. (Elvsåshagen et al., 2012) did not observe a change in the C1 component. In the present study, P1 was also significantly increased between post-modulation and pre-modulation in the control group, although there were no significant overall group of time effects or interactions effects for C1.
We did not find a significant difference between groups in modulation of the N1 component. In contrast to previous studies, we did observe a decreased N1 amplitude in both groups during continuous visual stimulation, but this was in the opposite direction of N1 modulation observed in previous studies (Normann et al., 2007;Elvsåshagen et al., 2012;Zak et al., 2018). In these studies, a modulation effect of N1 was shown as an increase of the N1 amplitude. The opposite direction of N1 modulation may be due to the high variability observed in the N1 latency and amplitude. In contrast to the latencies of the C1 and P1 components, the latency of the N1 component had a wide range of 150-200 ms post-stimulus (Fig. 3). Previous studies often used distinct protocols to identify the N1 component or it remained unmentioned.
The largest cohort of control subjects (n=415) involved in measuring VEP plasticity after 10 minutes of continuous visual stimulation was reported by Valstad et al. (Valstad et al., 2020).
They observed a strong modulation effect with decreased C1, and increased P1, N1 and N1b amplitudes 2-6 min after modulation. In the present study, we have not focused on the N1b component, but on the early VEP components in accordance with Elvsåshagen et al. and Normann et al. (Normann et al., 2007;Elvsåshagen et al., 2012). Early VEP components might be less affected by attention or complex cognitive processes (Normann et al., 2007). Attention may especially explain the variation in N1 amplitude (Luck et al., 2000). Future studies should take the subject's attention carefully into account by implementation of an additional attention test (Kloppel et al., 2015;Valstad et al., 2020).
Our observed duration of the potentiated VEP response in controls appeared to be shorter than some of previous studies. Normann et al. (Normann et al., 2007) showed changes in VEP amplitudes after continuous visual stimulation up to 20 min, although in some individual experiments the VEP amplitudes were potentiated up to 60 min. In the present study, the strongest modulation is observed during the 10-min modulation block and 2 min postmodulation. The latter is consistent with the findings of Valstad et al. (Valstad et al., 2020). The duration of the VEP response could depend on a variety of factors, including high inter-subject variability in the VEP response, degree of neural recruitment, and level of attention to the visual stimulus. Increasing the duration of post-modulation recording could have allowed for modulation effects later in time (i.e. >30 min), which is following the definition of LTP (Normann et al., 2007;Lisman, 2017). Notably, we choose our methodology due to its feasibility in patients, prior demonstration of a robust modulation effect in large cohorts of unaffected controls and the lack of a VEP response in psychiatric patients (Normann et al., 2007;Valstad et al., 2020).
To our knowledge, the present study is the first to examine VEP potentiation during the modulation block, which revealed a potentiated VEP response during continuous visual stimulation without delay in both groups. In animal studies, it has been shown that an increased response during LTP induction enhances the response after induction (Kushner et al., 2005).
But although VEP plasticity shows similarities to the properties of synaptic plasticity, it is unknown whether the potentiation during prolonged visual stimulation is dependent upon synaptic plasticity (Cooke and Bliss, 2006).

VEP plasticity in NF1
In contrast to control subjects, we observed no potentiation of the VEP response of the P1 component during continuous visual stimulation and post-modulation in adults with NF1.
These findings support the theory of decreased synaptic plasticity found in animal models of NF1 (Costa et al., 2002;Cui et al., 2008;Shilyansky et al., 2010;Omrani et al., 2015). These studies describe that reduced NF1 activity in animal models of NF1 leads to an increase in gamma-aminobutyric acid (GABA) neurotransmission, which causes a decrease in glutamatergic synaptic plasticity. In support of this theory, previous studies in adults with NF1 showed alterations in motor cortical excitability and plasticity upon a form of repetitive transcranial magnetic stimulation (Mainberger et al., 2013;Castricum et al., 2020).
Furthermore, magnetic resonance spectroscopy showed reduced GABA levels in the visual cortex of NF1 patients (Violante et al., 2013(Violante et al., , 2016 that could affect cortical excitability and plasticity. Interestingly, there was a potentiated VEP response during continuous visual stimulation without delay in the NF1 group and control group. This could indicate that a normal response in VEP potentiation was observed during continuous visual stimulation in the NF1 group, but the ability to maintain this potentiation was affected post-modulation in the NF1 group. A similar effect has been observed in a previous study on motor cortical plasticity using transcranial magnetic stimulation in NF1 patients (Castricum et al., 2020). These previous studies are in line with the present study, which together indicates that adults with NF1 may have reduced visual cortical plasticity, as indicated by their non-potentiated response following VEP induction.
Interestingly, as mentioned before, the N1 modulation effect in both groups was in the opposing direction as reported in previous studies, and showed high inter-subject variability in latency and amplitude, which makes the interpretation of the difference more difficult.
Increasing the number of electrodes to record the VEP could decrease variability. Future studies of NF1 patients should be performed to further characterize the modulation effect of N1. In addition, the strength of the cortical visual plasticity could depend on the history of neural activity, i.e. metaplasticity in the human visual cortex (Abraham and Bear, 1996;Bocci et al., 2014). Metaplasticity affects the ability to induce subsequent long-lasting synaptic plasticity (Abraham and Bear, 1996). In future research, metaplasticity should be taken into account by, for example, standardizing the history of neural activity using a combination of neurophysiological techniques such as TMS or transcranial direct current stimulation, and EEG (Bocci et al., 2014;Opie et al., 2017). Furthermore, future studies should investigate whether metaplasticity is aberant in patients with NF1 as has been found in several psychiatric and neurological disorders (Müller-Dahlhaus and Ziemann, 2015).
An important limitation of the study is that our NF1 sample may not be representative.
In our NF1 sample, there may have been a participation bias towards highly motivated or less cognitively affected patients. The differences in VEP potentiation between NF1 and controls may have been larger in a more severe cognitively affected NF1 sample. Furthermore, it has been mentioned that VEP components are influenced by attention, although early components might be less affected (Normann et al., 2007). Attention in NF1 patients may be reduced due to fatigue. Increased fatigue has been associated with NF1 and has been shown to affect the daily life of adults with NF1 (Rietman et al., 2018). However, we did not find any difference in fatigue based on a sleepiness scale at the start or end of the VEP recordings. Nevertheless, we acknowledge the lack of control for eye movements during stimulation. We did instruct and motivate the subjects to keep their focus during the experiment. Furthermore, the used binocular pattern stimulation facilitates fixation and attention (Odom et al., 2010). In addition, the VEP amplitude of P1 usually varies minimally with repeated recordings (Odom et al., 2004).
Nevertheless, future studies may benefit from the inclusion of a motivational feature to keep focus during the experiment, such as the implementation of a reward during the experiment.
The implementation of an additional test during the non-invasive stimulation paradigms, such as reading out numbers or pressing a key with changing colors (Kloppel et al., 2015;Valstad et al., 2020), could improve the subject's attention. Overall, we did not observe any deviations in the VEP recordings that could indicate visual deprivation.
Although we did not find a significant modulation effect in the P1 amplitude of the VEP in the NF1 group, the results should be considered cautiously due to the relatively small sample size of the NF1 group, and the small effect sizes. The strengths of the study were that the NF1 patients were not receiving mental health care and were not using psychoactive medication.
Additionally, the control and NF1 groups were similar in age and sex, and the experiment was standardized to the time of day. Hence, these factors could not explain the differences in potentiation between the NF1 and control groups.
In conclusion, we showed that VEP plasticity can be measured in response to prolonged stimulation of low frequency checkerboard reversals. The non-potentiated response upon VEP modulation in patients with NF1 may indicate deficits in visual cortical plasticity. Due to the small NF1 sample, small effect sizes, and transient potentiation in controls, the results should be considered with caution. Future studies should investigate VEP plasticity more extensively by including a longer period of post-modulation and studying late VEP components in a larger group of patients in which attentional measures are considered. The present study contributes to an improved assessment of the feasibility for using neurophysiological outcome measures in intervention studies of cognitive deficits among patients with NF1. Table 1. Demographics (Mean ± SD), and VEP responses pre-modulation, post-modulation, and during modulation of the neurofibromatosis type 1 (NF1) group and the control group separately.