The organophosphorus compound soman, an irreversible inhibitor of cholinesterases, produces seizure activity and related brain damage. Studies using various biochemical markers of programmed cell death (PCD) suggested that soman-induced cell damage in the brain was apoptotic rather than necrotic. However, it has recently become clear that not all PCD is apoptotic, and the unequivocal demonstration of apoptosis requires ultrastructural examination. Therefore, the present study was undertaken to reinvestigate the damage produced in the brains of mice sacrificed at various times within the first 24 h or at 7 days after a convulsive dose of soman. Classical histology and ultrastructural examination were performed. The immunohistochemical expression of proteins (p53, Bax) involved in PCD, DNA fragmentation (TUNEL method at light and electron microscopy levels) and the glial reaction were also explored. Our study confirms that the severity of lesions depended on the duration of convulsions and shows that cerebral changes were still occurring as late as 7 days after the onset of long-lasting convulsions. Our observations also establish that there was a large variety of ultrastructurally distinct types of cell damage, including hybrid forms between apoptosis and necrosis, but that pure apoptosis was very rare. A prominent expression of p53 and Bax proteins was detected indicating that PCD mechanisms were certainly involved in the morphologically diverse forms of cell death. Since purely apoptotic cells were very rare, these protein expressions were presumably involved either in nonapoptotic cell death mechanisms or in apoptotic mechanisms occurring in parallel with nonapoptotic ones. Moreover, evidence for DNA fragmentation by the TUNEL method was found in apoptotic but also in numerous other morphotypes of cell damage. Therefore, TUNEL-positivity and the expression of PCD-related proteins, in the absence of ultrastructural confirmation, were here shown not to provide proof of apoptosis. In soman poisoning as well as in other cerebral pathologies, premature conclusions on this question can potentially be misleading and might even lead to detrimental therapies.