Developmental effectsLower complexity scales and age

Significance level for correlations was Bonferroni-corrected to p < 0.017, accounting for the average correlation between the eight ROIs [38]. In FXS participants, positive correlations between lower complexity scales and age were found in Cz (r = 0.36, p = 0.015), FCz (r = 0.39, p = 0.007), Oz (r = 0.41, p = 0.004), Pz (r = 0.43, p = 0.003), TL (r = 0.37, p = 0.012), and TR (r = 0.37, p = 0.012). In ASD participants, positive correlations were found across almost all ROIs (Cz: r = 0.49, p < 0.001; FL: r = 0.37, p = 0.010; FCz: r = 0.50, p < 0.001; Oz: r = 0.56, p < 0.001; Pz: r = 0.57, p < 0.001; TL: r = 0.52, p < 0.001; TR: r = 0.53, p < 0.001) except FR (r = 0.34, p = 0.019), where the correlation did not survive correction. In NT controls, lower complexity scales positively correlated with age in all ROIs (Cz: r = 0.52, p < 0.001; FL: r = 0.41, p = 0.004; FR: r = 0.43, p = 0.002; FCz: r = 0.53, p < 0.001; Oz: r = 0.57, p < 0.001; Pz: r = 0.59, p < 0.001; TL: r = 0.53, p < 0.001; TR: r = 0.54, p < 0.001).

Higher complexity scales and age

Significance level for correlations was corrected to p < 0.007. In FXS participants, negative correlations were found in FL (r = -0.31; p = 0.036) and FR (r = -0.31; p = 0.04), but they did not survive correction for multiple correlations. In ASD participants and NT controls, no correlations were found.

APF and age

Significance level for correlations was corrected to p < 0.017. In FXS participants, APF did not correlate with age in any ROI. In ASD participants, positive correlations were found in FCz (r = 0.37; p = 0.010), and TR (r = 0.38; p < 0.008). Positive correlations were also found in Cz (r = 0.33, p = 0.022), FL (r = 0.31, p = 0.035), FR (r = 0.32, p = 0.027), Pz (r = 0.34, p = 0.02), and TL (r = 0.33, p = 0.021), but they did not survive correction. Lastly, in NT controls, positive correlations with age were found in all ROIs (Cz: r = 0.44, p = 0.001; FL: r = 0.53, p < 0.001; FR: r = 0.44, p = 0.001; FCz: r = 0.45, p = 0.001; Oz: r = 0.37, p = 0.007; Pz: r = 0.48, p < 0.001; TL: r = 0.50, p < 0.001; TR: r = 0.46, p < 0.001).

Delta and age

Significance level for correlations was corrected to p < 0.016. In FXS participants, negative correlations with age were found across almost all ROIs (Cz: r = -0.50, p < 0.001; FL: r = -0.60, p < 0.001; FCz: r = -0.55, p < 0.001; Oz: r = -0.49, p < 0.001; Pz: r = -0.46, p = 0.001; TL: r = -0.49, p < 0.001; TR: r = -0.50, p < 0.001) except FR (r = -0.34; p = 0.022), where the correlation did not survive correction. In ASD participants, delta correlated negatively with age in Cz (r = -0.46, p = 0.001), FCz (r = -0.40, p = 0.005), Oz (r = -0.42, p = 0.003), TL (r = -0.58; p < 0.009), and TR (r = -0.38, p = 0.008). Negative correlations were also found in FR (r = -0.34; p = 0.017) and Pz (r = -0.30; p = 0.036), but they did not survive correction for multiple correlations. In NT controls, negative correlations with age were found in almost all ROIs (Cz: r = -0.48, p < 0.001; FR: r = -0.40, p = 0.004; FCz: r = -0.55, p < 0.001; Oz: r = -0.34, p = 0.014; Pz: r = -0.47, p < 0.001; TL: r = -0.59, p < 0.001; TR: r = -0.38, p = 0.015) except FL (r = -0.18; p = 0.21).

Theta and age

Significance level for correlations was corrected to p < 0.009. In FXS participants, a negative correlation with age was found in FL (r = -0.41; p = 0.004). Negative correlations were also found in Cz (r = -0.36; p = 0.017) and Oz (r = -0.36; p = 0.014), but they did not survive correction. In ASD participants, theta did not correlate with age in any ROI. In NT controls, theta correlated negatively with age in FR (r = -0.44; p = 0.001), and Pz (r = -0.43; p = 0.002). Negative correlations were also found in Cz (r = -0.36; p = 0.009), FCz (r = -0.36; p = 0.011), and TL (r = -0.35; p = 0.012), but they did not survive correction for multiple correlations.

Alpha and age

Significance level for correlations was corrected to p < 0.009. In FXS participants, alpha did not correlate with age in any ROI. In ASD participants, positive correlations were found across almost all ROIs (Cz: r = 0.35, p = 0.016; FR: r = 0.33, p = 0.023; FL: r = 0.32, p = 0.029; FCz: r = 0.37, p = 0.01; Oz: r = 0.38, p = 0.009; Pz: r = 0.36, p = 0.014; TR: r = 0.34, p = 0.022) except TL (r = 0.21; p = 0.16), but they did not survive correction for multiple correlations. In NT controls, alpha correlated positively with age in Cz (r = 0.46, p < 0.001), FR (r = 0.39, p = 0.006), FCz (r = 0.53, p < 0.001), Oz (r = 0.43, p = 0.002), Pz (r = 0.52, p < 0.001), and TL (r = 0.53, p < 0.001). Positive correlations were also found in FL (r = 0.36; p = 0.011) and TR (r = 0.35; p = 0.014), but they did not survive correction.

Beta and age

Significance level for correlations was corrected to p < 0.014. In FXS participants, beta correlated positively with age in FCz (r = 0.39; p = 0.008), Pz (r = 0.37; p = 0.011), TL (r = 0.48; p = 0.001), and TR (r = 0.46; p = 0.002). In ASD participants, positive correlations with age were found in Cz, (r = 0.42; p = 0.003), FCz (r = 0.53; p < 0.001), Oz (r = 0.45; p = 0.001), and TL (r = 0.51; p < 0.001). Positive correlations were also found in Pz (r = 0.35; p = 0.014) and TR (r = 0.33; p = 0.023), but they did not survive correction. In NT controls, beta correlated positively with age in Cz (r = 0.36; p = 0.009), FCz (r = 0.43; p = 0.002), TL (r = 0.45; p < 0.001), and TR (r = 0.47; p < 0.001). A positive correlation was also found in Pz (r = 0.29; p = 0.04), but it did not survive correction for multiple correlations.

Low gamma and age

Significance level for correlations was corrected to p < 0.012. In FXS participants, low gamma only correlated positively with age in Pz (r = 0.39; p = 0.007). Positive correlations were also found in FCz (r = 0.31; p = 0.027) and TL (r = 0.34; p = 0.021), but they did not survive correction. In ASD participants, low gamma did not correlate with age in any ROI, although positive correlations were found in Cz (r = 0.31; p = 0.033) and TL (r = 0.30; p = 0.04), but did not survive correction. In NT controls, a positive correlation with age was found in TL (r = 0.36; p = 0.010). Positive correlations were also found in FL (r = 0.35; p = 0.012) and TR (r = 0.35; p = 0.013), but they did not survive correction.

High gamma and age

Significance level for correlations was corrected to p < 0.012. In FXS participants, positive correlations with age were found across almost all ROIs (Cz: r = 0.50, p < 0.001; FL: r = 0.41, p = 0.006; FCz: r = 0.58, p < 0.001; Oz: r = 0.55, p < 0.001; Pz: r = 0.56, p < 0.001; TL: r = 0.42, p = 0.005; TR: r = 0.44, p = 0.002) except FR (r = 0.36; p = 0.017), which did not survive correction. However, in ASD participants, high gamma did not correlate with age in any ROI. In NT controls, high gamma only correlated positively with age in FL (r = 0.36; p = 0.010).

Resting state markersSignal complexity

Figure 2 shows the MSE curve in the central (A) and frontal right (B) regions for the three groups.

Fig. 2

MSE curve in Cz (A) and FR (B) for FXS (orange), ASD (purple), and NT controls (blue)

Lower and higher complexity scales

Results showed a significant Scales*Group interaction (F(2,135) = 14.47, p < 0.001, η2 = 0.18), suggesting that values in the lower and higher scales differed between the three groups. Further analyses were performed to identify group differences in lower and higher complexity scales.

Lower complexity scales (S1-20)

MANCOVA with age as covariate showed significant effects of Age (F(8,128) = 13.03, p < 0.001, η2 = 0.45) and Group (F(16,256) = 4.29, p < 0.001, η2 = 0.21). Bonferroni-corrected post hoc analyses revealed reduced complexity in lower scales in NT controls compared to ASD participants in all ROIs (Cz: p = 0.002; FL: p = 0.019; FR: p = 0.004; FCz: p = 0.002; Oz: p = 0.005; Pz: p = 0.005; TL: p = 0.002; TR: p = 0.014), and diminished complexity in lower scales in FXS participants compared to ASD participants in Cz (p = 0.009), FCz (p = 0.009), Oz (p = 0.031), Pz (p = 0.043), and TL (p = 0.025). No differences were found between FXS participants and NT controls. MANCOVA with age and NVIQ as covariates showed significant effects of Age (F(8,127) = 12.10, p < 0.001, η2 = 0.43) and Group (F(16,254) = 4.18, p < 0.001, η2 = 0.21), but no significant effect of NVIQ (F(8,127) = 1.64, p = 0.12, η2 = 0.09). Post hoc analyses with Bonferroni corrections showed decreased complexity in lower scales in NT controls compared to ASD participants in all ROIs (Cz: p = 0.003; FL: p = 0.013; FR: p = 0.005; FCz: p < 0.002; Oz: p = 0.008; Pz: p = 0.008; TL: p = 0.003; and TR: p = 0.017) when controlling for NVIQ. However, no differences were found between FXS and ASD participants, and again, there were no differences between FXS participants and NT controls.

Higher complexity scales (S21-40)

Results showed a significant effect of Group (F(2,126) = 6.15, p = 0.003, η2 = 0.09). Bonferroni-corrected follow-up ANOVA revealed reduced complexity in higher scales in FXS participants compared to NT controls in all ROIs (Cz: p = 0.002; FL: p = 0.025; FR: p = 0.024; FCz: p = 0.005; Oz: p < 0.001; Pz: p < 0.001; TL: p = 0.015; TR: p < 0.001), and compared to ASD participants in all ROIs (Cz: p = 0.005; FL: p = 0.020; FR: p = 0.004; FCz: p = 0.042; Oz: p = 0.019; Pz: p = 0.004; TR: p = 0.018) except TL (p = 0.07). No differences were found between ASD participants and NT controls. Results with NVIQ as covariate showed no effects of NVIQ (F(1,125) = 0.00, p = 0.99, η2 = 0.00) or Group (F(2,125) = 2.19, p = 0.12, η2 = 0.03), and no differences between the groups were found in any ROI.

Power spectral density

Figure 3 shows the average power spectra in the central (A) and temporal right regions (B) for the three groups.

Fig. 3

Group average spectra in Cz (A) and TR (B) for FXS (orange), ASD (purple), and NT controls (blue)

APF

MANCOVA with age as covariate showed significant effects of Age (F(8,130) = 2.60, p = 0.012, η2 = 0.14) and Group (F(16,260) = 1.80, p = 0.031, η2 = 0.10). Bonferroni-corrected post hoc analyses revealed lower APF in FXS participants compared to NT controls in all ROIs (Cz: p < 0.001; FL: p = 0.002; FR: p < 0.001; FCz: p < 0.001; Oz: p = 0.005; Pz: p = 0.002; TL: p < 0.001; TR: p = 0.001), as well as lower APF in FXS compared to ASD participants in Cz (p = 0.002), FL (p = 0.048), FR (p = 0.01), FCz (p = 0.005), Pz (p = 0.031), and TR (p = 0.035). No differences were found between ASD participants and NT controls. Figure 4 shows differences in APF in Cz between the three groups. MANCOVA with age and NVIQ as covariates only revealed a main effect of Age (F(8,129) = 3.01, p = 0.004, η2 = 0.16). No NVIQ (F(8,129) = 1.11, p = 0.36, η2 = 0.06) or Group (F(16,258) = 0.82, p = 0.66, η2 = 0.05) effects were observed, and no differences between the groups were found in any ROI.

Fig. 4

APF in Cz between FXS (orange), ASD (purple), and controls (blue). Dotted lines indicate the median for each group

Delta

MANCOVA with age as covariate showed a significant effect of Age (F(8,121) = 5.82, p < 0.001, η2 = 0.27), but no effect of Group (F(16,242) = 0.65, p = 0.83, η2 = 0.04). No differences were found between the groups in any ROI. MANCOVA with age and NVIQ as covariates showed a significant effect of Age (F(8,120) = 5.05, p < 0.001, η2 = 0.25), but no effects of NVIQ (F(8,120) = 0.59, p = 0.78, η2 = 0.04) or Group (F(16,240) = 0.60, p = 0.87, η2 = 0.04) were observed. Again, no differences were found between the groups in any ROI.

Theta

MANCOVA with age as covariate showed a significant effect of Group (F(16,136) = 4.31, p < 0.001, η2 = 0.33), but no effect of Age (F(8,118) = 1.61, p = 0.13, η2 = 0.10). Follow-up analyses with Bonferroni corrections revealed higher theta power in FXS compared to NT controls in all ROIs (Cz: p = 0.012; FL: p = 0.001; FR: p < 0.001; FCz: p = 0.007; Oz: p = 0.027; Pz: p < 0.001; TL: p = 0.041) except TR (p = 1.00), and compared to ASD participants in FR (p = 0.004), FCz (p < 0.001), Pz (p = 0.015), TL (p < 0.001), and TR (p = 0.005). NT controls showed higher theta power compared to ASD participants in TL (p = 0.009) and TR (p = 0.006), but these differences did not reach the threshold. MANCOVA with age and NVIQ as covariates showed a significant effect of Group (F(16,234) = 3.01, p < 0.001, η2 = 0.17), but no effects of Age (F(8,117) = 1.56, p = 0.14, η2 = 0.10) or NVIQ (F(8,117) = 0.63, p = 0.75, η2 = 0.04) were observed. Bonferroni-corrected follow-up analyses showed increased theta power in FXS participants compared to ASD participants in FCz (p = 0.002) and TL (p = 0.003), as well as higher theta power in NT controls compared to ASD participants in TL (p = 0.004) and TR (p = 0.002), but these differences did not reach 25% of scalp coverage.

Alpha

MANCOVA with age as covariate showed significant effects of Age (F(8,121) = 2.96, p = 0.005, η2 = 0.16) and Group (F(16,242) = 3.90, p < 0.001, η2 = 0.21). Follow-up analyses with Bonferroni corrections revealed lower alpha power in FXS participants compared to NT controls in all ROIs (Cz: p < 0.001; FL: p < 0.001; FR: p < 0.001; FCz: p < 0.001; Oz: p < 0.001; Pz: p = 0.003; TL: p < 0.001; TR: p < 0.001). FXS participants also showed lower alpha power compared to ASD participants in Cz (p = 0.012), FL (p = 0.001), FR (p = 0.01), Oz (p < 0.001), Pz (p = 0.004), and TR (p = 0.032). ASD participants showed reduced alpha power compared to NT controls in TL (p = 0.022) and TR (p = 0.001), but these differences did not reach the threshold. MANCOVA with age and NVIQ as covariates showed significant effects of Age (F(8,122) = 2.98, p = 0.004, η2 = 0.17) and Group (F(16,240) = 2.76, p < 0.001, η2 = 0.16), but no effect of NVIQ (F(8,120) = 0.76, p = 0.64, η2 = 0.05) was found. Bonferroni-corrected follow-up analyses showed reduced alpha power in FXS participants compared to NT controls in TL (p = 0.004) and TR (p < 0.001), and compared to ASD participants in FL (p = 0.033) and Oz (p = 0.015). Lower alpha power was also observed in ASD participants compared to NT controls in TL (p = 0.038) and TR (p = 0.004). However, none of these differences reached the threshold of 25% of scalp coverage.

Beta

MANCOVA with age as covariate showed significant effects of Age (F(8,117) = 6.93, p < 0.001, η2 = 0.32) and Group (F(16,234) = 2.32, p = 0.003, η2 = 0.14). Follow-up analyses with Bonferroni corrections revealed higher beta power in ASD participants compared to NT controls in all ROIs (Cz: p = 0.002; FL: p < 0.001; FR: p < 0.001; FCz: p < 0.001; Oz: p = 0.003; Pz: p < 0.001; TL: p = 0.003; TR: p < 0.001). Beta power was also higher in ASD participants compared to FXS participants in FL (p = 0.002), FR (p = 0.015), and Oz (p = 0.024). No differences were found between FXS participants and NT controls. MANCOVA with age and NVIQ as covariates showed significant effects of Age (F(8,116) = 5.97, p < 0.001, η2 = 0.29) and Group (F(16,232) = 1.81, p = 0.031, η2 = 0.11), but no effect of NVIQ (F(8,116) = 0.89, p = 0.52, η2 = 0.06) was observed. Follow-up analyses with Bonferroni corrections showed higher beta power in ASD participants compared to NT controls in all ROIs (Cz: p = 0.004; FL: p < 0.001; FR: p < 0.001; FCz: p < 0.001; Oz: p = 0.002; Pz: p < 0.001; TL: p = 0.008; TR: p < 0.001) when controlling for NVIQ. No differences were found between FXS participants and NT controls, or between FXS and ASD participants.

Low gamma

MANCOVA with age as covariate showed significant effects of Age (F(8,118) = 2.49, p = 0.016, η2 = 0.14) and Group (F(16,236) = 4.20, p < 0.001, η2 = 0.22). Bonferroni-corrected post hoc analyses revealed exaggerated low gamma power in FXS participants compared to NT controls in Cz (p < 0.001), FR (p = 0.03), FCz (p < 0.001), Pz (p = 0.031), TL (p < 0.001), and TR (p < 0.001), as well as higher low gamma power in ASD participants compared to NT controls in all ROIs (Cz: p = 0.014; FL: p = 0.009; FR: p = 0.031; FCz: p < 0.001; Pz: p = 0.019; TL: p < 0.001; TR: p < 0.001) except Oz (p = 1.00). No differences were found between FXS and ASD participants. Figure 5 shows differences in low gamma power in FCz between the three groups. MANCOVA with age and NVIQ as covariates showed main effects of Age (F(8,117) = 2.36, p = 0.021, η2 = 0.14) and Group (F(16,234) = 2.94, p < 0.001, η2 = 0.17), but no effect of NVIQ (F(8,117) = 0.13, p = 1.00, η2 = 0.01) was found. Post hoc analyses with Bonferroni corrections showed exaggerated low gamma power in FXS participants compared to NT controls in Cz (p = 0.006), FCz (p < 0.001), TL (p = 0.005), and TR (p = 0.002) when controlling for NVIQ. Increased low gamma power was also observed in ASD participants compared to NT controls in all ROIs (Cz: p = 0.012; FL: p = 0.007; FR: p = 0.33; FCz: p < 0.001; Pz: p = 0.018; TL: p < 0.001; TR: p < 0.001) except Oz (p = 1.00). Again, no differences were found between FXS and ASD participants.

Fig. 5

Low gamma power in FCz between FXS (orange), ASD (purple), and NT controls (blue). Dotted lines indicate the median for each group

High gamma

MANCOVA with age as covariate showed a significant effect of Group (F(16,240) = 3.72, p < 0.001, η2 = 0.20), but no effect of Age (F(8,120) = 1.52, p = 0.16, η2 = 0.09). Bonferroni-corrected post hoc analyses showed exaggerated high gamma power in ASD participants compared to NT controls in Cz (p = 0.009), FL (p = 0.043), FCz (p = 0.001), TL (p < 0.001), and TR (p < 0.001). Increased high gamma power was observed in ASD participants compared to FXS participants in FL (p = 0.016) and TL (p = 0.017), but these differences did not reach the threshold. No differences were found between FXS participants and NT controls. MANCOVA with age and NVIQ as covariates only showed a main effect of Group (F(16,238) = 3.09, p < 0.001, η2 = 0.17), but no effects of Age (F(8,119) = 1.28, p = 0.26, η2 = 0.08) or NVIQ (F(8,119) = 0.24, p = 0.98, η2 = 0.02) were found. Bonferroni-corrected post hoc analyses showed exaggerated high gamma power in ASD participants compared to NT controls in Cz (p = 0.024), FCz (p = 0.007), TL (p < 0.001), and TR (p < 0.001) when controlling for NVIQ. High gamma power was also higher in ASD participants compared to FXS participants in FL (p = 0.045) and TL (p = 0.018), but it did not reach the 25% of scalp coverage. Again, no differences were found between FXS participants and NT controls.