Many experiments showed that neurons tend to fire at some phase (usually trough) of local field potential (LFP) oscillations, such as theta or gamma rhythm. LFP is supposedly generated by a population of neurons with coherent currents induced by spikes. So there is a causal link from spiking activity to LFPs. Is there also a significant influence on the spikes directly from LFP (not via the hidden spikes that generated the LFP)?
Is it just an epiphenomenon, and or is there a possibility that it is partly a serious mechanism for neural computation?