This paper describes a method for detecting epileptic spikes in a record electroencephalographic (EEG) surface by taking a single channel. We identified a pattern using multiresolution analysis with a biorthogonal wavelet after processing and analyzing the Wavelet Toolbox of Matlab, 207 records and 132 records of tips tricks previously classified by Neurophysiologist. This pattern enabled an algorithm for detecting spikes in patients with refractory temporal lobe epilepsy, based on the maximum voltage in each of the six levels of reconstruction using biorthogonal 3.7 wavelet. The algorithm was applied on records of patients with epilepsy, getting a sensitivity of 92% and a specificity of 80% in the diagnosis of epileptic spikes.