Background: While many biological studies can be performed on cell-based systems, the investigation of molecular pathways related to complex human dysfunctions - e.g. neurodegenerative diseases - often requires long-term studies in animal models. The nematode Caenorhabditis elegans represents one of the best model organisms for many of these tests and, therefore, versatile and automated systems for accurate time-resolved analyses on C. elegans are becoming highly desirable tools in the field. Results: We describe a new multi-functional platform for C. elegans analytical research, enabling automated worm isolation and culture, reversible worm immobilization and long-term high-resolution imaging, and this under active control of the main culture parameters, including temperature. We employ our platform for in vivo observation of biomolecules and automated analysis of protein aggregation in a C. elegans model for amyotrophic lateral sclerosis (ALS). Our device allows monitoring the growth rate and development of each worm, at single animal resolution, within a matrix of microfluidic chambers. We demonstrate the progression of individual protein aggregates, i.e. mutated human superoxide dismutase 1 - Yellow Fluorescent Protein (SOD1-YFP) fusion proteins in the body wall muscles, for each worm and over several days. Moreover, by combining reversible worm immobilization and on-chip high-resolution imaging, our method allows precisely localizing the expression of biomolecules within the worms' tissues, as well as monitoring the evolution of single aggregates over consecutive days at the sub-cellular level. We also show the suitability of our system for protein aggregation monitoring in a C. elegans Huntington disease (HD) model, and demonstrate the system's ability to study long-term doxycycline treatment-linked modification of protein aggregation profiles in the ALS model. Conclusion: Our microfluidic-based method allows analyzing in vivo the long-term dynamics of protein aggregation phenomena in C. elegans at unprecedented resolution. Pharmacological screenings on neurodegenerative disease C. elegans models may strongly benefit from this method in the near future, because of its full automation and high-throughput potential.