Long QT syndrome (LQTS) is an inherited disorder associated with prolongation of the QT/QTc interval on the surface electrocardiogram (ECG) and a markedly increased risk of sudden cardiac death due to cardiac arrhythmias. Up to 25% of genotype-positive LQTS patients have QT/QTc intervals in the normal range. These patients are, however, still at increased risk of life-threatening events compared to their genotype-negative siblings. Previous studies have shown that analysis of T-wave morphology may enhance discrimination between control and LQTS patients. In this study we tested the hypothesis that automated analysis of T-wave morphology from Holter ECG recordings could distinguish between control and LQTS patients with QTc values in the range 400-450 ms. Holter ECGs were obtained from the Telemetric and Holter ECG Warehouse (THEW) database. Frequency binned averaged ECG waveforms were obtained and extracted T-waves were fitted with a combination of 3 sigmoid functions (upslope, downslope and switch) or two 9th order polynomial functions (upslope and downslope). Neural network classifiers, based on parameters obtained from the sigmoid or polynomial fits to the 1 Hz and 1.3 Hz ECG waveforms, were able to achieve up to 92% discrimination between control and LQTS patients and 88% discrimination between LQTS1 and LQTS2 patients. When we analysed a subgroup of subjects with normal QT intervals (400-450 ms, 67 controls and 61 LQTS), T-wave morphology based parameters enabled 90% discrimination between control and LQTS patients, compared to only 71% when the groups were classified based on QTc alone. In summary, our Holter ECG analysis algorithms demonstrate the feasibility of using automated analysis of T-wave morphology to distinguish LQTS patients, even those with normal QTc, from healthy controls.