Atypical processing of tones and phonemes in Rett Syndrome as biomarkers of disease progression

Rett Syndrome (RTT), caused by mutations in the MECP2 gene on the X chromosome, presents with severe motor and cognitive impairments and typically minimal or absent expressive language, making behavioral assessment of auditory processing difficult. High-density EEG recordings of auditory evoked potentials (AEPs) offer an objective way to probe auditory cortical processing and potentially serve as biomarkers of disease severity, progression, and treatment response. Prior auditory brainstem response (ABR) studies suggest largely intact subcortical processing, whereas cortical AEP studies indicate substantial impairments and slowed processing in RTT, including prolonged latencies (Pa, N1, P2) and atypical development of N2. Recent work shows delayed mismatch negativity (MMN) to pitch changes and impaired auditory sensory memory for duration in RTT. Speech-evoked AEPs in RTT have shown abnormalities at later latencies (circa 200–750 ms), but early components (P1, N1, P2, N2) to speech have not been systematically examined due to limited trial counts in prior studies. This study probes the integrity of successive early cortical auditory processing stages to both simple tones and complex phonemes in females with RTT to determine where along the hierarchy deficits emerge and to evaluate biomarkers of disease severity.

- ABR findings in RTT generally indicate preserved early subcortical processing (e.g., normal ABR), but cortical AEPs reveal significant impairments. - Reported cortical abnormalities include prolonged latencies of Pa, N1, P2 and atypical developmental changes in N2 latency. - MMN studies indicate delayed cortical representation of basic auditory features such as frequency and impaired auditory sensory memory for duration; MMN to duration was present only at rapid stimulation rates (~2 Hz) and absent at slower rates (~0.5–1 Hz). - Speech AEP studies (e.g., responses to real words vs non-words and to own-name vs other names) revealed abnormalities in RTT at 200–750 ms but did not resolve earlier components due to few trials. - Early AEP components (P1, N1, P2, N2) are sensitive to stimulus complexity and implicated in speech processing in children, underscoring the need to systematically study these components in RTT using sufficient trial counts.

Design: Passive oddball EEG study with two auditory conditions: simple tones and phonemes. Analyses focused on frequent standard stimuli to ensure high signal-to-noise for early AEP components. Participants: 13 female RTT patients (mean age 12.9 years; range 3.9–20.6); diagnosis confirmed clinically and genetically; symptom severity assessed with Rett Syndrome Severity Scale (RSSS). One RTT participant excluded due to excessive EEG noise. Control group: 21 age-matched typically developing (TD) females (mean age 12.44; range 4.3–21.1) with no neurodevelopmental/psychiatric history; normal hearing. Groups were age-matched (t(31)=0.32, p=0.75). Experimental setup: Participants sat in a sound-attenuated, electrically shielded booth, watched a silent movie to maintain wakefulness. Auditory stimuli presented via speakers. Oddball paradigm with standards (p=0.85) and deviants (p=0.15); analyses reported here focus on standards. Stimuli: - Tone condition: 1000 Hz standard; 500 Hz deviant; 100 ms duration; 10 ms rise/fall; 75 dB SPL; SOA 900 ms. - Phoneme condition: /ba/ standard; /da/ deviant; female speaker; 250 ms duration; 65 dB SPL; SOA 900 ms. Blocks: RTT completed ~9.6 blocks per condition (range 7–11); TD ~10 blocks (range 10–11); 140 stimuli per block; total recording ~40 min; breaks provided. Block order (tones first for most) did not affect AEPs in TD as per ANOVA. EEG acquisition: Biosemi ActiveTwo 64-channel system; sampling 512 Hz; DC–150 Hz passband during acquisition; common mode sense/driven right leg reference setup. Preprocessing: Epochs from −200 to +600 ms around stimulus onset; bandpass 1–20 Hz (zero-phase Butterworth) to focus on early cortical components and reduce muscle noise; artifact rejection by z-score/variability (trials >2 SD removed); bad channels interpolated. One RTT excluded for excessive noise. Mean retained trials: Tone RTT 756, TD 804; Phoneme RTT 733, TD 763 (range 558–906). Baseline −200 to 0 ms; re-referenced to average of TP7/TP8. Analyses focused on FCz/FC3/FC4 where auditory AEPs are maximal. AEP components and statistics: Component windows P1 (60–90 ms), N1 (100–130 ms), P2 (135–165 ms), N2 (245–275 ms). Mean amplitudes in these windows entered into mixed ANOVAs: Group (RTT vs TD) × Stimulus-type (Tone vs Phoneme). Supplementary peak-based latency/amplitude analyses in Tone condition used automated peak detection (P1 ≥40 ms first positive; N1 80–200 ms first negative; P2 next positive; N2 200–400 ms most negative). Some subjects without identifiable peaks were excluded from peak analyses. Additional analyses: ROC to assess classification performance and optimal cut-offs; split-half reliability (odd/even trials) with Spearman-Brown correction; Pearson correlations with age and, within RTT, P2 vs RSSS; Fisher z to compare age–AEP correlations across groups.

- P1: Significant Stimulus-type × Group interaction. In TD, P1 amplitude was larger for phonemes than tones (t(20)=5.185, p<0.0001). This modulation by stimulus complexity was absent in RTT (t(11)=0.046, p=0.964). No main Group effect on P1 amplitude. - N1: Significant main effect of Stimulus-type (reduced amplitude for phonemes vs tones) across both groups; no Group effect or interaction. - P2: Markedly reduced in RTT regardless of stimulus. Main effect of Group: F=36.928, p<0.0001, η2=0.544; no Stimulus-type effect or interaction. Group averages indicate near absence of P2 in RTT. Example means (μV ± SD): Tone P2 TD 3.20±1.73 vs RTT −0.72±1.50; Phoneme P2 TD 2.94±2.03 vs RTT 0.29±1.92. - N2: Smaller in RTT with main Group effect: F=10.756, p=0.003, η2=0.258; no Stimulus-type effect. Example means (μV ± SD): Tone N2 TD −3.30±2.41 vs RTT −0.62±2.43; Phoneme N2 TD −2.87±1.83 vs RTT −1.14±1.72. - Robust differentiation: P2 provided near-perfect separation between RTT and TD across a wide age range; only 3 RTT participants exceeded the minimum TD P2 value, and still below TD mean. - Clinical relevance: P2 amplitude (averaged across conditions) correlated with RTT symptom severity (RSSS): r(12)=−0.62, p=0.032 (lower P2 associated with greater severity). - Additional: Grand-averaged AEPs in TD showed clear P1, N1, P2, N2 to both tones and phonemes; RTT deviated markedly, with only P1 consistently discernible.

The study set out to characterize early cortical auditory processing in RTT and to identify potential electrophysiological biomarkers. The findings show that while the earliest cortical response (P1) is present in RTT, the normal enhancement of P1 for complex phonemic stimuli seen in TD peers is absent, indicating an early deficit in sensitivity to stimulus complexity. The N1 appears relatively preserved across groups, suggesting that some early cortical encoding mechanisms remain intact in RTT. In stark contrast, the mid-latency P2 is profoundly diminished or absent in RTT across both tones and phonemes, and N2 is also significantly reduced. These results indicate prominent deficits at successive stages beyond the initial cortical encoding, impacting integration or higher-order processing of auditory inputs. The robustness of the P2 reduction, its ability to separate RTT from TD across ages, and its negative association with clinical severity (RSSS) support the P2 amplitude as a promising objective biomarker for monitoring disease status, progression, and treatment response. The similarity of P2 attenuation to findings in RTT animal models further underscores its translational relevance and potential utility for bridging preclinical and clinical studies.

This study provides a detailed characterization of early auditory cortical processing in females with RTT using high-density AEPs to tones and phonemes. Key contributions include: (1) identification of absent P1 modulation by stimulus complexity in RTT, (2) preservation of N1, and (3) profound, stimulus-independent reductions of P2 (and reduced N2), with P2 strongly differentiating RTT from TD and correlating with clinical severity. These findings position the P2 amplitude as a compelling candidate biomarker for RTT that may serve in monitoring, treatment response assessment, or as a surrogate endpoint, and offer a translational link to animal models. Future work should evaluate longitudinal changes, validate P2 as a clinical trial biomarker, and extend analyses to additional auditory metrics (e.g., higher-frequency responses such as ASSR) and broader speech processing paradigms.

- Sample size was modest, and one RTT participant was excluded due to excessive EEG noise; some individuals lacked clearly detectable peaks for supplementary latency analyses. - The main windowed-amplitude analysis is not sensitive to latency shifts; although supplementary peak analyses were conducted, details are limited in the present text. - A restricted 1–20 Hz bandpass was used to mitigate muscle noise in RTT, limiting assessment of higher-frequency activity. - Block order effects were evaluated only in TD participants; potential order effects in RTT remain untested. - Use of passive listening with standards-focused analyses may not capture all aspects of speech and auditory processing relevant to communication in RTT.