Visual evoked potentialsDemographics and clinical variables

Demographics and clinical severity scores for participants included in the analysis of the VEP are provided in Table 1. There was a significant group difference in CSS (H(3) = 9.65, p = 0.022) with follow-up pairwise comparisons indicating greater severity for the CDD group compared to the MDS group (p = 0.005). There were no significant group differences in MBA (H(3) = 7.38, p = 0.061). There was a significant group difference in seizure frequency (H(3) = 27.0, p < 0.001) with greater seizure frequency in participants with CDD compared to participants with RTT (p < 0.001) and MDS (p < 0.001).

Group comparison

Kruskal–Wallis H tests indicated significant group differences in VEP N1, N1–P1, and P1–N2 amplitude (H(4) = 15.3, p = 0.004; H(4) = 24.4, p < 0.001; H(4) = 25.1 p < 0.001, respectively; Fig. 1A-B). Follow-up comparisons indicated reduced N1 amplitude in participants with RTT compared to TD participants (p < 0.001), reduced N1–P1 amplitude in participants with RTT (p < 0.001), CDD (p = 0.001), and MDS (p = 0.003) compared to TD participants, and reduced P1–N2 amplitude in participants with RTT (p < 0.001) and MDS (p < 0.001) compared to TD participants. There were no significant differences in VEP amplitude between the FOXG1 group and the other groups or significant differences in VEP latency for any of the groups (Table 3).

Fig. 1

Visual Evoked Potentials. A Grand average VEP waveforms for TD, RTT, CDD, MDS, and FOXG1 groups at electrode Oz. B Box plots showing the distribution of the latency and amplitude of the VEP components for TD, RTT, CDD, MDS, and FOXG1 participants. The amplitude of multiple components was reduced in RTT, CDD, and MDS groups compared to the TD group. Statistical analyses were performed with Kruskal–Wallis tests and post-hoc tests analyses with Bonferroni-adjusted p values for significance (*p < 0.005). C Scatterplots showing the association between VEP N1-P1 amplitude and clinical severity (CSS and MBA) for RTT, CDD, MDS, and FOXG1 participants. N1-P1 amplitude was significantly associated with both severity measures in participants with RTT and FOXG1. No aspects of the VEP were significantly associated with severity for the other groups. Statistical analyses were performed using linear regression (*p < 0.05)

Table 3 Visual evoked potential latency and amplitudeAssociations with clinical severity: VEP amplitude

In participants with RTT, VEP N1 and N1-P2 amplitudes were significantly associated with clinical severity with decreasing amplitude with increasing severity (N1 amplitude & CSS: R2 = 0.096, F (1, 42) = 4.34, β = -0.309, p = 0.044; N1-P1 amplitude & CSS: R2 = 0.103, F (1, 42) = 4.72, β = -0.321, p = 0.036; N1-P1 amplitude & MBA: R2 = 0.120, F (1, 42) = 5.62, β = -0.347, p = 0.023). VEP N1-P2 amplitude was also negatively associated with CSS in participants with FOXG1 (R2 = 0.782, F (1, 4) = 10.77, β = -0.884, p = 0.046), however this association was largely driven by a single participant with milder symptoms (see Fig. 1C). Clinical severity using the current measures was not significantly associated with any of the VEP parameters for participants with CDD or MDS (Fig. 1C). Additional analyses were conducted to determine if the absence of a P1 was associated with greater clinical severity. This analysis demonstrated no difference in CSS or MBA score for participants who were excluded from the analysis of the VEP for the absence of a P1 component (n = 22), on average, compared to those participants with an identifiable P1 (n = 79; CSS: p = 0.133; MBA: p = 0.169).

Associations with age

VEP N1-P1 and P1-N2 amplitudes decreased with age in participants with RTT (R2 = 0.115, β = -0.338, p = 0.026; R2 = 0.129, β = -0.359, p = 0.018, respectively). There were no significant associations between age and the VEP for the TD, CDD, MDS, or FOXG1 groups (Fig. 3).

Auditory evoked potentialsDemographics and clinical variables

Demographics and severity scores for participants included in the analysis of the AEP are provided in Table 2. There were no significant group differences in overall severity on the CSS or MBA (H(3) = 4.42, p = 0.220; H(3) = 4.88, p = 0.181, respectively). There was a significant group difference in seizure frequency (H(3) = 20.9, p < 0.001) with pairwise comparisons indicating greater seizure frequency in CDD compared to RTT (p < 0.001) and MDS (p = 0.001).

Group comparison

There was a significant effect of group on AEP P1, N1, and P2 latency (H(4) = 15.4, p = 0.004; H(4) = 26.5, p < 0.001; H(4) = 11.1, p = 0.011, respectively; Fig. 2A). Follow-up pairwise comparisons indicated delayed P1 latency in MDS compared to RTT (p < 0.001), delayed N1 latency in MDS compared to RTT (p < 0.001) and CDD (p < 0.001), delayed N1 latency in FOXG1 compared to CDD (p = 0.003) and RTT (p = 0.005), and delayed P2 latency in MDS compared to RTT (p < 0.001). There were no group differences in the amplitude of the AEP components (Table 4).

Fig. 2

Auditory Evoked Potentials. A Box plots showing the distribution of the latency and amplitude of the AEP components for TD, RTT, CDD, MDS, and FOXG1 participants. The peak latencies of AEP components were prolonged in participants with MDS and FOXG1 compared to participants with RTT and CDD. Statistical analyses were performed with Kruskal–Wallis tests and post-hoc analyses with Bonferroni-adjusted p values for significance (*p < 0.005). B Scatterplots illustrating the association between AEP P1-N1 amplitude and the severity measures (CSS and MBA) for RTT, CDD, MDS, and FOXG1 participants. P1-N1 amplitude was significantly associated with severity in RTT and CDD. C Scatterplots showing the association between the latency of select AEP components and the clinical severity measures in participants with RTT, CDD, MDS, and FOXG1. The latency of one or more of the AEP components was significantly associated with severity in CDD, MDS, and FOXG1. Statistical analyses for B-C were performed using linear regression (*p < 0.05). The grand average AEP waveforms are not included due to age-related shifts in peak latency, particularly for the TD group, which obscure comparisons of peak amplitudes

Table 4 Auditory evoked potential latency and amplitudeAssociations with clinical severity: AEP amplitude

In participants with RTT, AEP P1 and P1-N1 amplitudes were associated with CSS with decreasing amplitude with increasing severity (P1 amplitude: R2 = 0.123, F (1, 44) = 6.89, β = -0.351, p = 0.012; P1 – N1 amplitude: R2 = 0.089, F (1, 50) = 4.78, β = -0.298, p = 0.034). AEP amplitudes were also negatively associated with severity in participants with CDD. Specifically, in participants with CDD, P1 and P1–N1 amplitudes were associated with both CSS and MBA (CSS: P1 amplitude: R2 = 0.455, F (1, 13) = 10.03, β = -0.675, p = 0.008; P1–N1 amplitude: R2 = 0.487, F (1, 13) = 11.41, β = -0.698, p = 0.005; MBA: P1 amplitude: R2 = 0.353, F (1, 13) = 6.53, β = -0.594, p = 0.025; P1–N1 amplitude: R2 = 0.380, F (1, 13) = 7.37, β = -0.617, p = 0.019). The association between N1-P2 amplitude and severity was specific to the CSS (R2 = 0.380, F (1, 13) = 7.36, β = -0.617, p = 0.019). There were no associations between AEP amplitude and clinical severity for participants with MDS or FOXG1 (Fig. 2B). Additional analyses were conducted to determine if the absence of an AEP N1 peak was associated with clinical severity. This analysis revealed that participants without a N1 component (n = 14) had more severe CSS, on average, compared to those with a N1 (n = 85; p < 0.001, η2 = 0.109). There was no group difference in MBA scores based on the presence or absence of a N1 (p = 0.116).

Associations with clinical severity: AEP latency

In participants with CDD, N1 and P2 latencies were associated with CSS with decreasing latency with increasing severity (N1 latency: R2 = 0.470, F (1, 13) = 10.64, β = -0.698, p = 0.007; P2 latency: R2 = 0.322, F (1, 13) = 5.69, β = -0.567, p = 0.034; see Fig. 2C). P1 latency was also negatively associated with severity in participants with MDS (CSS: R2 = 0.360, F (1, 13) = 6.74, β = -0.600, p = 0.023; MBA: R2 = 0.388, F (1, 13) = 7.60, β = -0.623, p = 0.017). In participants with FOXG1, P2 latency was associated with MBA, with increasing latency with increasing severity (R2 = 0.741, F (1, 5) = 9.48, β = 0.861, p = 0.028; Fig. 2C). There was no association of latency with severity in the RTT cohort.

Associations with age

The latency of the AEP components declined with age in TD participants in line with the established literature on the typical maturation of the AEP (P1 latency: R2 = 0.420, F (1, 22) = 15.2, β = -0.648, p < 0.001; N1 latency: R2 = 0.579, F (1, 22) = 28.9, β = -0.761, p < 0.001; P2 latency,: R2 = 0.295, F (1, 22) = 13.7, β = -0.544 p = 0.007). AEP P1-N1 amplitude also decreased with age in TD participants (R2 = 0.179, F (1, 22) = 4.59, β = -0.423, p = 0.044). There were no associations between age and the AEP in participants with RTT, CDD, MDS, or FOXG1 (Fig. 3).

Fig. 3

Associations between AEP and VEP Latency and Age for All Groups. A Scatterplot of AEP N1 latency and age in TD, RTT, CDD, MDS, and FOXG1 participants. N1 latency declined significantly with age in TD participants. Age was not significantly associated with AEP latency for participants with RTT, CDD, MDS, or FOXG1. B Scatterplot of VEP P1 latency and age for all groups. VEP latency did not change with age in any of the groups, consistent with the established literature on the stability of VEP P1 latency from early childhood. Statistical analyses for were performed using linear regression (*p < 0.05)