The method of lesion analysis is a tool widely used in cognitive neuroscience to conduct group studies with stroke patients associating cognitive dysfunctions with brain lesions. Although modern statistical lesion analysis (voxel wise lesion symptom mapping, VLSM) is an established tool there is still a lively debate about the time between stroke onset, neuroimaging and the assessment of cognitive functions. There are several reasons supporting acute/subacute or chronic neuroimaging or functional assessment. Performing lesion mapping in acute stages avoids the influences of uncontrollable variables such as individual rehabilitative interventions and resources obviously biasing data collection in the chronic phase. Further, deficits occurring only in the acute phase will be overlooked. On the other hand, high resolution images at a chronic stage when acute consequences of brain lesions, such as tissue swellings, have decreased represent a more accurate method of lesion localization than qualitative inferior clinical imaging. The present study aimed to investigate how post-stroke time of structural imaging and functional assessment may influence the results of lesion studies. We compared the results of anatomical and clinical data in acute versus chronic stages after stroke. Therefore, 55 patients were scanned in acute ( 4 months post stroke) stages after stroke with various neuroimaging techniques (CT, T2FLAIR, DWI, T1). As a functional measure motor deficits were assessed using BMRC (British medical research council) scores of (upper) limbs in acute and chronic stages. With established knowledge about the neuronal substrate of motor functions we were able to assess the validity of either method - acute and chronic lesion mapping. A subsample of patients ( n = 30) was scanned in the chronic phase with high-resolution MR techniques (T1, T2FLAIR), while from the other patients only CT or clinical MR scans were available. Normalized lesion maps from every patient and time point were compared intraindividually. Additionally, for the subsample of patients with high-resolution chronic imaging comparisons between T1 and T2FLAIR scans were conducted. For all conditions (chronic vs. acute, T1 vs. T2FLAIR) VLSM analyses were performed and the results were compared to known anatomy of motor functions. A preliminary data analysis demonstrated that acute and chronic/T1 and T2FLAIR scans differed significantly: chronic lesion maps drawn on T2FLAIR images tended to overestimate, chronic T1 scans to underestimate acute lesion size. We further observed obvious mislocalizations of chronic periventricular lesions through projection of the normalized lesion map onto a standard brain template. The VLSM analyses revealed different functional results depending on time point and image modality. However, the acute lesion approach showed the most accurate mapping of motor functions in the brain. We suspect that typical morphological changes over time cause a regional bias of chronic lesion maps. Especially regions in close vicinity to the ventricles may cause vast errors in lesion localization. Further, it is to assume that deficits associated with a lesion decreased/vanished due to regular neuronal reorganization processes or rehabilitative interventions. This clearly biased the measurement of behavioral in the chronic phase. These results suggest a preferential use of acute structural neuroimaging data for the method of lesion analysis.