Abstracts

Rationale: Blood oxygenation level-dependent (BOLD) contrast is commonly used for functional MRI (fMRI) studies. The temporal evolution of BOLD signal changes is in the form of a characteristic curve called the hemodynamic response function (HRF). Since the BOLD signal arises from local changes in blood flow, blood volume and oxidative metabolism, it is possible that the HRF varies with underlying abnormalities of brain tissue. Recently, an increasing number of studies have applied EEG-fMRI to epilepsy. However, it is unknown whether HRF parameters are consistent in different lesions. In this study, we examined epilepsy patients with focal cortical dysplasia (FCD) and those with mesial temporal sclerosis (MTS) and compare their HRF parameters.Methods: Twenty seven focal epilepsy patients who had been diagnosed with FCD (n=14) or unilateral MTS (n=13) were selected from our EEG-fMRI database. EEG was recorded inside 3T MRI scanner with 25 electrodes placed on the scalp according to 10-20 system. For each patient, epileptic discharges similar to those recorded outside the scanner were marked as events. The timing of events was convolved with HRFs at multiple latencies. A statistical t-map was created showing voxels which significantly correlated with marked EEG events. We computed the HRF at voxels with maximum t-value (t>|3.1|) selected from activated and deactivated regions. We measured the amplitude and latency of the HRF peak (Fig. 1) and compared the FCD and MTS groups using t-test with a threshold of 0.05. To compare the two groups, we analyzed 1) HRF in peak activation, 2) HRF in peak deactivation, 3) maximum absolute response within the EEG spike field, activation or deactivation.Results: Nineteen patients showed both activation and deactivation; 6, activation only; 2, deactivation only. Fig 2 shows the HRF peak amplitude and latency in FCD and MTS groups. HRF amplitude in deactivation was significantly larger in MTS than in FCD. Looking at this result in detail, 5 MTS patients had deactivation in default mode network (DMN) regions and their amplitude was larger than that of 5 FCD patients with deactivation in DMN. HRF amplitude in activation and HRF latency in activation and deactivation were not significantly different between the two groups. There was also no significant difference between the two groups when analyzing the dominant response from each patient.Conclusions: We found that HRF peak amplitude in deactivated region was larger in MTS than in FCD and this difference is prominent in DMN regions. Our results suggest that mesial temporal spikes have more effect on the DMN than spikes in other regions. However, if we focus on the dominant response, there is no major difference between FCD and MTS. The current study indicated that it is not necessary to use different HRF models for EEG-fMRI studies in patients with FCD and MTS. This study was supported by CIHR grant MOP-38079.