BACKGROUND: Analysis of intracranial ictal patterns may help to predict surgical outcomes. We investigated intracranial EEG patterns to correlate with surgical outcomes and compared the yield of 'subdural electrodes alone (SE)' versus 'combined depth and subdural electrodes (CDSE)' for ictal lateralization in temporal lobe epilepsy (TLE). METHODS: We reviewed a total of 95 seizures recorded by bilateral temporal depth and subdural electrodes in 25 TLE patients who underwent surgery. We classified surgical outcomes as 'seizure-free' or 'not-seizure-free'. Each seizure was analyzed based on the presence or absence of peri-ictal discharges, ictal distribution, waveform patterns, onset frequency and involved number of electrodes, and interhemispheric propagation time (IHPT). The accuracy of lateralizing seizure foci by CDSE was compared to that by SE. RESULTS: 20 patients (80.0%) were 'seizure-free' and 5 (20.0%) were 'not-seizure-free'. The presence of peri-ictal discharges (p<0.001), distribution of depth only or depth and medial electrodes (p<0.001) and higher onset frequency (p=0.021) were associated with 'seizure-free' outcomes. Ictal onset pattern with fast spike trains was common in 'seizure-free', whereas pattern with rhythmic activity was common in 'not-seizure-free' (p=0.005). SE correctly lateralized in 18 of 20 patients, and incorrectly lateralized in the remaining 2 patients, but CDSE correctly lateralized in all 20 patients. CONCLUSIONS: Some intracranial ictal patterns were significantly correlated with good surgical outcomes. These findings suggest that the analysis of ictal EEG patterns help to predict surgical outcomes. CDSE is more accurate for the lateralization of seizure foci compared to SE.