Lafora disease (LD) is an autosomal recessive, always fatal progressive myoclonus epilepsy with rapid cognitive and neurologic deterioration. One of the pathological hallmarks of LD is the presence of cytoplasmic PAS+polyglucosan inclusions called Lafora bodies (LBs). Current clinical and neuropathological views consider LBs to be the cause of neurological derangement of patients. A systematic study of the ontogeny and structural features of the LBs has not been done in the past. Therefore, we undertook a detailed microscopic analysis of the neuropile of a Laforin-deficient (epm2a-/-) mouse model. Wild type and epm2a-/- mice were sacrificed at different ages and their encephalon processed for light microscopy. Luxol-fast-blue, PAS, Bielschowski techniques, as well as immunocytochemistry (TUNEL, Caspase-3, Apaf-1, Cytochrome-C and Neurofilament L antibodies) were used. Young null mice (11 days old) showed necrotic neuronal death in the absence of LBs. Both cell death and LBs showed a progressive increment in size and number with age. Type I LBs emerged at two weeks of age and were distributed in somata and neurites. Type II LBs appeared around the second month of age and always showed a complex architecture and restricted to neuronal somata. Their number was considerably less than type I LBs. Bielschowski method showed neurofibrillary degeneration and senile-like plaques. These changes were more prominent in the hippocampus and ventral pons. Neurofibrillary tangles were already present in 11 days-old experimental animals, whereas senile-like plaques appeared around the third to fourth month of life. The encephalon of null mice was not uniformly affected: Diencephalic structures were spared, whereas cerebral cortex, basal ganglia, pons, hippocampus and cerebellum were notoriously affected. This uneven distribution was present even within the same structure, i.e., hippocampal sectors. Of special relevance, was the observation of the presence of immunoreactivity to neurofilament L on the external rim of type II LBs. Perhaps, type II LB is not the end point of a metabolic abnormality. Instead, we suggest that type II LB is a highly specialized structural and functional entity that emerges as a neuronal response to major carbohydrate metabolism impairment. Early necrotic cell death, neurocytoskeleton derangement, different structural and probably functional profiles for both forms of LBs, a potential relationship between the external rim of the LB type II and the cytoskeleton and an uneven distribution of these abnormalities indicate that LD is both a complex neurodegenerative disease and a glycogen metabolism disorder. Our findings are critical for future studies on disease mechanisms and therapies for LD. Interestingly, the neurodegenerative changes observed in this LD model can also be useful for understanding the process of dementia.