Abstracts

Rationale: Recognizing faces is an important brain function for social interaction. Impairment of face recognition without loss of visual acuity is called prosopagnosia and causes problems in recognizing a familiar or even the own face. Prosopagnosia can be a functional deficit caused by epilepsy or brain tumor surgery. Detailed functional mapping improves the functional outcome, but especially the gold standard electrical cortical stimulation (ECS) is hardly performed on the ventral temporal cortex, where face-selective cortical areas are located. Methods: In this study, six patients with epilepsy were candidates for epilepsy surgery and underwent a multimodal mapping procedure including ECS, electrocorticography (ECoG) and functional magnetic resonance imaging (fMRI). During ECS, the patients were stimulated while looking at faces and objects, and asked to report any changes. The ECoG and fMRI mappings detected significant high-γ activity or BOLD changes in response to faces compared to objects. The fMRI results within 6mm were mapped to the ECoG locations and considered as fMRI positive locations. Results: Face selective locations were found on the fusiform or inferior temporal gyrus in 4/6 patients in ECoG, followed by 3/6 in ECS and 2/6 in fMRI, as shown in Figure 1. In pairwise comparisons, ECoG and fMRI showed accuracies of 79.5% and 79.6% against ECS, whereas fMRI against ECoG resulted in 89.6% accuracy. In total, 176 subdural electrode locations were included in all mappings. Thus, 48 face selective locations were identified by at least one modality, showing a detection rate for ECoG, ECS and fMRI mapping of 73%, 44%, and 50%, respectively. A subset of 26 locations was confirmed by two of the modalities and 6 locations were covered by all of them. The detection rates of those locations identified by at least two modalities were 85%, 54%, 85% for ECoG, ECS and fMRI, respectively. Conclusions: This study showed that ECoG high-γ mapping is most sensitive to reveal face selective cortex regions. ECoG results were either confirmed by fMRI, ECS or even both. One case showed only ECoG positive and another one only ECS positive locations. These finding are supported by the fact that it does not rely on the patient’s ability to explain the symptoms during ECS and multiple ECoG locations can be tested in parallel.  In principle, fMRI has similar benefits compared to ECoG and also the 89.6% accuracy indicates that results overlap in both procedures. However, only in one case ECS and fMRI revealed overlapping positive locations. The variability of the mapping results across patients and modalities suggests that none of the modalities should be spared. For example, fMRI can help to reveal functional regions not covered by ECoG. In conclusion, high-γ mapping with ECoG turned out to be the most sensitive and reliable marker to identify face-selective regions. Funding: No funding