To understand the functional heterogeneity in neuronal activity associated with neurodevelopmental disorders (NDDs) and identify its molecular basis, we integrated RNA-sequencing and micro-electrode array (MEA) recordings in hiPSCs-derived neurons carrying NDD-related mutations. Clustering analysis of 345 recordings from 35 NDD-related cell lines at 28 days in-vitro (DIVs) revealed seven discrete activity phenotypes. While a continuous phenotypic landscape of network alterations emerged, specific clusters were identified for certain disorders. At DIV49, stronger associations between activity phenotypes and syndromes were observed. Simultaneous RNA-sequencing and network activity recording at DIV49 were performed for seven cell lines, including four lines with variants in chromatin remodeling genes (ADNP, YY1, CHD8 and EHMT1). Weighted Gene Correlation Network Analysis identified co-regulated gene modules potentially influencing the activity patterns. We built a Bayesian Network to integrate information on the expression level of each gene module and the functional clusters allocated to the samples, finding three parent nodes of the “cluster” variable that differentially combined to modulate the assigned clusters. Since the expression activity of these three parent gene modules could be influenced to induce particular network phenotypes and eventually to rescue disease-related activity patterns toward a control phenotype, we explored the Library of Integrated Network-based Cellular Signatures (LINCS) database for compounds enhancing the transcriptomic signature of specific clusters. In conclusion, our framework provides a platform to bridge the gap between molecular and functional aspects in normal and abnormal neurodevelopment, holding the potential to contribute to the development of treatments for NDDs.