Hi Taras,

Interesting question. I don't know the answer for any of the specific cases you cite, but from first principles we should be able to make some pretty good guesses.

Assuming simple neural properties, the local field potential goes negative when the charge density drops (i.e. positive charge carriers move into the cell). This is why the classical fEPSP is a negative going event, as when AMPA channels open, the predominant charge carrier is Na+ moving into the cell.

So in order for us to observe the positive going after potential you are talking about, positive charge must be moving out of the cell. This could also be negative charge moving into the cell, e.g. Cl- moving into the cell due to inhibition. However, I believe that theory is unlikely, as the Cl- reversal potential is so close to rest that it generates only small transmembrane currents, which means it will not significantly effect the total charge density (though this is still a reasonable hypothesis).

However, we should take a brief moment to consider what is happening INSIDE the cell: When the EPSP hits the cell, it is depolarized by cations flux into the cell. However, does it stay depolarized? No! It hyperpolarizes again, back to the exact same potential it had before the EPSP. Because neural membranes are capacitors, and hence follow the basic principle of V = Q/C... that means the to return to the same resting potential after an EPSP, an equal and opposite charge must flow onto the capacitor. Simply put the same amount of charge the moved into the cell during the EPSP flows back out of the cell afterward. If this is the cause of the positive going portion of the wave, then this means that the area under positive going wave and area under the negative going portions of the voltage transient should be equal and opposite. So I extracted the data from the second pulse in the 2hour wash EPSP in figure 1. The negative going portion of the EPSP had an area of -3.8 mV.ms and the positive going portion had an area of 3.5 mV.ms: pretty close to equal and opposite. Supporting my suggestion that we are simply seeing the charge transfer responsible for returning the cell back to the resting membrane potential.

So if the positive going portion of the wave is due to the charge transfer responsible for returning the cell to rest, how come we don't see it in every example trace? I don't know. My first guess is filtering: a lot of people perform high pass filtering when recordings fEPSPs, and the slow positive going potential may be removed. But I don't think that is the real cause. The other guess is because cells are electrical cables. All of the rough maths above relies on the notion that cells are isopotential spheres, which is plainly not the case. Because neurons are electronically large, and have different conductances throughout the membrane, we see a lot of behaviour that could not be explained by the assumptions I made above, i.e. that the fEPSP is positive going in the cell body layer, while the population spike is negative going, while in stratum radiatum, the inverse is true. 

In fact, we know that at the site of EPSP initiation, we do not get enough charge transfer after the EPSP to repolarize the entire cell. This charge transfer is spread throughout most of the cell, focused around the soma, but also distal to the site of EPSP initiation. Indeed, this is why you tend you see a bigger positive going portion of the wave if your fEPSP electrode is nearer to the soma, or right at the furthest reaches of CA1 cell dendrites. I suggest you have a read of some of the many paper that use current source density to analyze field events in the hippocampus, e.g. 

J Neurophysiol. 2001 Nov;86(5):2435-44.

Apical and basal orthodromic population spikes in hippocampal CA1 in vivo show differentorigins and patterns of propagation.

Kloosterman F1, Peloquin P, Leung LS.

J Neurophysiol. 1987 Nov;58(5):981-96.

Action-potential discharge in hippocampal CA1 pyramidal neurons: current source-densityanalysis.

Richardson TL, Turner RW, Miller JJ.

So, long story short, the positive going part of an EPSP is most likely due to the charge transfer needed to repolarize the cell. You are more likely to see that when your recording electrode  are somewhat closer to the soma, or right at end end of CA1 dendrites. Have larger or small fEPSP electrodes will also alter this, as large electrodes will take the average of the field potential over the conductive area of the electrode tip.

I hope that was of some help, and of some interest.

Yes, it may be an issue of stim and recording electrode placement -  alignment in st. radiatum or distance between electrodes, which differs systematically from the studies you refer to.  Other factors are EPSP-IPSP sequences - feedforward IPSPs. 

Cheers

Hi Taras,

William has already given you an excellent and in-depth explanation for the likely source of this positive-going potential (pun intended). I would agree the most likely cause is electrode placement. If you are recording in CA1, try placing both stim and recording electrodes as close as possible to the 'middle' of the stratum radiatum (i.e. equidistant from the cell body later and the S.L.M.), which should help to minimise this effect. Here's another paper examining current source density in CA1, which helped me to better understand this phenomenon: https://www.ncbi.nlm.nih.gov/pubmed/8930359

Cheers,

Patrick.

https://www.ncbi.nlm.nih.gov/pubmed/8930359

Great, thanks all for the valuable discussion! I'll try to vary the position of the electrodes in a wider range than I did before and see what happens with fEPSP shape.

I suspect it is caused by di-synaptic (or polysynaptic) feedback inhibition. Because feedback inhibition takes time to take effect, that may explain why this positive deflection has a longer latency. Try different levels of stimulations and see if this postive deflection would disappear with a reduced stim intensity. Another way is to perform your fEPSP recording while blokcing all inhibitory transmission. If blocking inhibitory transmission abolishes the positive deflection in your fEPSP, it must be caused by GABAergic transmission.