Abstract
BACKGROUND: Cochlear implants (CIs) restore hearing by directly stimulating spiral ganglion neurons (SGNs), yet auditory outcomes remain highly variable. Increasing evidence suggests that SGN survival alone incompletely predicts CI performance; instead, transcriptional programmes governing neuronal excitability/synaptic transmission, structural plasticity, trophic-metabolic support and injury/inflammation may better reflect neural functional competence. METHODS: We re-analysed publicly available cochlear transcriptomic datasets spanning development, adulthood and injury/degeneration. Primary resources included developmental FACS RNA-seq of hair cells and surrounding tissue, adult inner- and outer-hair-cell microarrays, and a noise-induced hearing loss (NIHL) RNA-seq cohort following mesenchymal stromal cell (MSC) therapy. We additionally analysed independent injury/degeneration datasets, including spatial transcriptomics of spiral ganglion regions after noise exposure and spiral ganglion RNA-seq after aminoglycoside-induced deafening. Guided by cochlear neuroscience literature and functional enrichment, we assembled gene modules for excitability, plasticity, trophic/metabolic support and injury/inflammation. Module activity was quantified using within-dataset standardized scores from normalized expression, avoiding cross-platform merging. We derived a Cochlear Neural Functional Competence (CNFC) score (Excitability + Trophic - Injury) and assessed robustness using a minimal 24-gene panel. External validation was performed in an independent purified SGN dataset, and CNFC was benchmarked against transcriptome-wide principal components. RESULTS: Developmental maturation was characterized by increasing excitability-associated transcripts alongside down-regulation of actin/cytoskeletal remodelling components. Adult hair cells displayed distinct trophic signatures. In the NIHL model, MSC therapy was associated with transcriptional suppression of excitatory receptor and channel genes, consistent with a shift in the injury/inflammation-excitability balance, although functional consequences remain to be established. Importantly, in independent injury/degeneration datasets, CNFC decreased in spiral ganglion neuronal regions after noise exposure and in deafened spiral ganglion. Across datasets, CNFC captured coherent trends and remained highly correlated with full-module scoring when reduced to the 24-gene panel. CONCLUSION: CNFC is a transparent, hypothesis-generating framework for summarizing cochlear neuronal functional state from transcriptomic data, complementing traditional survival metrics. By prioritizing interpretable modules and standardized within-dataset scoring, CNFC supports cross-study integration and highlights candidate programmes for mechanistic testing.