Abstracts

Rationale: Complex visuo-spatial functional MRI (fMRI) tasks reliably activate posterior mesial temporal lobe (hippocampus-parahippocampus) bilaterally in healthy controls. Studies on patients with hippocampal sclerosis and mesial temporal lobe epilepsy (mTLE) suggest lateralisation of visuo-spatial memory fMRI activation to the healthy side. We aimed to assess visuo-spatial memory in patients with mTLE and amygdala lesions compared to healthy controls. Methods: Twenty patients (9 women; median age 36 years, range 21-58) with mTLE due to unilateral lesions in amygdala and 19 healthy controls (10 women, median age 28 years, range 22-53) were tested with fMRI Roland's Hometown Walking Task: mental navigation and recall of familial visuo-spatial landmarks. All patients underwent neuropsychological assessment. Median age at seizure onset was 24 years, range 6-50, median epilepsy duration - 6 years, range 1-34. Seventeen patients (85%) had drug resistant seizures; 18 (90%) were right-handed. All patients underwent at least two high resolution MRI (1.5T) with an interval of at least six months for diagnostic purposes. Results: Bilateral symmetrical hippocampal fMRI activations were observed more frequently in healthy controls compared to patients: 68% vs. 35% (p=0.05). However, the rate of bilateral parahippocampal BOLD signal was similar in patients and controls: 85% vs. 95% (p=0.6). In patients, there was no statistically significant correlation between the lesion and BOLD signal sides. The patients with the left sided amygdala lesions (n=16) had either bilateral (7/16), ipsilateral (2/16) or contralateral (n=2/16) BOLD signal in hippocampi. There was no hippocampal activation in two patients. In the majority of patients with left-sided amygdala lesions, there was a bilateral activation of parahippocampal areas (13/16). Single patients had ipsi- and contralateral fMRI activations; one patient had no BOLD signal in parahippocampal area. The patients with right-sided amygdala lesions (n=4) had either ipsilateral hippocampal BOLD signal (2/4) or no hippocampal activation (2/4). Parahippocampal areas were activated bilaterally in all four patients with right-sided amygdala lesions. In neuropsychological testing, deficits in visual memory were seen only in 3/20 (15%) patients. All three had left-sided amygdala lesion and bilateral fMRI activation of hippocampal-parahippocampal areas. Conclusions: fMRI activation patterns elicited by Roland's Hometown Walking Task differ between patients with unilateral mTLE and healthy controls. Integration of epileptogenic lesions into non-verbal memory processing might have important implications for possible post-surgical deficits in patients with drug resistant mTLE. The study has been supported by the Austrian Science Fund (FWF); Project number P 21636-B18