This question is way too general - there are too many unknown parameters (what kind of information do you want to get out of it? what tools are you already familiar with?). There is a HUGE variety of techniques used for analysis of ephys data, and often EEG people have completely different approaches to those analyzing intracellular recordings - you can imagine that the challenges faced by each, and the type of data produced, are very different.

I would suggest that you post a separate question for each one of these techniques (LFP and EEG could be together), while also specifying the recording area, acquisition methods, and your existing background. Good luck

Good suggestions by Roni Hogri.....if you haven't yet suggest you start with a course on random data analysis, also a text such Random Data Analysis by Bendat & Piersol would be a good reference.

A good starting point can be the book: Analyzing Neural Time Series Data - Theory and Practice by Mike X Cohen.

Hello Rachel,

EEG, LFP and MUA belong to a family of signals called Gaussian signals that are often formed in nature when many contributing sources are involved (cells in your case) . In practice, this means that your recordings are fully characterized by the mean and (co)variance. In brain research, time zero is usually the stimulus time, and the averaging is done over multiple responses (each having its most recent stimulus as time zero). In EEG jargon, the mean is called ERP (Evoked Response Potential). When it comes to MUA, the mean is often zero due to filtering, but the average of the square of the signal (the variance) is not zero. Furthermore, if your electrodes are located in a region which is directly involved in the response to stimulus in question, then the mean and variance often have a characteristic shape.

Intracellular recordings capture just one cell and are therefore different from the family of signals mentioned above. If the cell is a spiking neuron, then you expect to see a train of spikes where each spike looks roughly the same but the frequency of their occurrences change in time. These signal are analyzed using a PSTH (peristimulus histogram, see wiki). The averaging and counting of the PSTH is again performed with respect to the nearest stimulus, and often a characteristic response appears, especially if the recorded brain region is directly involved in the response to the stimuli.

I hope you find this helpful,

Nir