'Independent component analysis' is a technique of data transformation that finds independent sources of activity in recorded mixtures of sources. It can be used to recover fluctuations of membrane potential from individual neurons in multiple-detector optical recordings. There are some examples in which more than 100 neurons can be separated simultaneously. Independent component analysis automatically separates overlapping action potentials, recovers action potentials of different sizes from the same neuron, removes artifacts and finds the position of each neuron on the detector array. One limitation is that the number of sources--neurons and artifacts--must be equal to or less than the number of simultaneous recordings. Independent component analysis also has many other applications in neuroscience including, removal of artifacts from EEG data, identification of spatially independent brain regions in fMRI recordings and determination of population codes in multi-unit recordings.