Miniature/spontaneous postsynaptic currents

Recording and analyzing miniature/spontaneous postsynaptic currents (m/sPSCs) is one of the most common experiments in patch clamp electrophysiology. m/sPSCs are ionic currents from AMPA, NMDA, glycine, or GABAA receptors that are evoked due to release of a single or multiple synaptic vesicles. We will cover both mPSCs and sPSCs, and go over how to analyze these small PSC events. While this chapter focuses on PSCs, most of the theory applies to miniature/spontaneous postsynaptic potentials (PSPs) as well. There are several experiments you can do that utilize PSCs; synapse number, silent synapses, excitatory/inhibitory ratio, synaptic multiplicity, and changes in synaptic release regulated by other non-ionic receptors.

Miniature postsynaptic currents (mPSCs)¶

Miniature postsynaptic currents (mPSCs) are ionic currents evoked from the release of a single synaptic vesicle Del Castillo & Katz, 1954. Frequency of mPSCs is used as a proxy for the number of functional synapses (synapses that have presynaptic input) that contain the receptor of interest (but not necessarily the number of synapses). The interpretation of mPSC data depends on what receptor you are recording from. If you are recording mEPSCs from AMPARs then you are likely getting the number of “active” or “non-silent” synapses (i.e. the number of presynaptic elements that have a postsynaptic elements with the receptor you are interested in). If you are recording NMDARs you will get the number of synapses with NMDAR receptors. If you are recording mIPSCs then you are getting the number of inhibitory synapses. With mIPSCs you could be getting GABAAR or GlyR unlees you block one or the other. One important caveat of mPSCs is that you are not getting where the presynaptic input is coming from. If you want projection specific synaptic input you need to run a different type of experiment than we will be covering in a later chapter.

Internal and external solutions¶

To record mEPSCs you will need to block spontaneous activity by including tetrodotoxin (TTX) in the bath. Preferably you would also use an internal solution that contains Cs+ (blocks potassium channels) and QX-314 (blocks voltaged-gated Na+ channels) to help with space clamp. For more information on internals see the internal solutions chapter. Depending on the receptor current(s) you want to record you will need to block other receptors by including specific drugs in the external solutions. For more information on externals see the external solutions chapter.

How should you record mPSCs¶

mPSCs are currents which means you are recording in voltage-clamp mode. This means that the amplifier will injecct current into the cell to keep it at your choosen holding voltage. Any changes in current means that the cells had a change in voltage due to a current that the amplifier is counteracting. When you get an inward current of positive ions the amplifier will inject a negative current to keep the cell at your chosen holding voltage.

There are two primary ways you can record mPSCs. One way is you can record continuously for about 3-5 minutes. Technically speaking this is the easiest method since you have a single recording and pretty much any simple recording software will implement this method. The second way is you can record 20-40 sweeps/acquisitions of 5-15 seconds each. Each of these acquisitions act as a kind of technical replicate. This method allows you to discard bad portions of the recording. Usually when you get proficient at patching you will rarely have bad recordings however sometimes you get 30 seconds where there is an unstable seal, digital cell phone noise, your bath gets too low, you get pump/vacuum noise or other issues. When this happens it is fine to discard the 5 or so acquisitions that are bad. You can create acquisitions from continuous recordings by splitting to get the same benefits of the sweeps/acquisitions method.

Filtering and sample rate are the other important considerations for capturing mPSCs. You generally want the sample rate to be 3-4x greater than the filter cutoff you are using with the filter cutoff determining the high frequency you are interested in. While signal theory says you have to have the sample rate 2x greater than the high frequency you are interested in, generally to capture that high frequency well you need to sample 3-4x times that rate to get a good resolution signal. In general mPSCs are recorded at a 10000 Hz with a 3000 Hz lowpass cutoff. A 10000 Hz sample rate is high enough to capture the rise of a mPSCs which are fairly quick, especially for mEPSCs on parvalbumin interneurons. 10000 Hz is also a good trade off between accuracy of the signal and digital storage space. In modern times you could realistically record at 20000 Hz with no storage space issues. I suggest a mininum sample rate of 10000 Hz.

Interpreting mPSCs¶

There are several important features of mPSCs that you will want to analyze. PSC frequency is used as a proxy for the number of synapses that contain the receptor whose currents you are recording. The more mPSCs, the more synapses. However, if there are changes in release probability you could also get a change in frequency without a change in synapse number. To determine whether there are changes in release probability you will need to run some pair-pulse experiments which are described in a later chapter. Another feature is PSC amplitude. A larger amplitude could mean two things. Larger mPSCs could mean there are more receptors at the postsynaptic element (thus a larger current). Alternatively, larger mPSC amplitude could mean that you get less distal synapses due to decreased dendritic length or more synapses close to the cell body with no change in dendritic length. More distal mPSCs should have a slower rise rate due to dendritic lowpass filtering. Lastly, you can look at the tau or decay rate of the mPSC. Changes in tau are usually due to changes in receptor subunit composition. Tau is especially useful when you need to identify specific cells types. Interneurons, like parvalbumin interneurons, have a very short mPESC tau compared to pyramidal neurons. Tau can also be affected by dendritic filtering.

Lastly, most of the data you get from mPSCs will be lognormal, exponential or gamma distributed. Rarely will you get normally distributed data. This has big implication for downstream analysis. For example the mean of a lognormal distribution is just the geometric mean: 10**np.mean(np.log10(data)). If your data is gamma distributed then log transforming with just skew the data in opposite direction which makes it much harder to get good measure of central tendency, however a square root transform also works well. If you want to learn more about distributions check out the distributions chapter.

Spontaneous postsynaptic currents (sPSCs)¶

Unlike mEPSCs, you do not want to include TTX in the bath since you are trying to get neurally evoked PSCs. You can use a Cs+ based internal with QX-314 though to improve space clamp. One reason to record sPSCs is to see how different receptors may change synaptic input to a cell by changing the activity/release probability of the presynaptic neurons/axons. sPSC experiments usually involve recording baseline sPSCS for 10 min, then flowing in your drug/peptide for 10 min and finally washing out the drug/peptide for 10 min. Changes in sPSC frequency suggest that the presynaptic neuron is affected by whatever drug you applied to the bath. If you just see a change in the holding current (the amount of current it takes to hold your cell at the mV you are holding it at) then there is likely a postsynaptic effect of the peptide.

Why do postsynaptic currents look they way they do?¶

Postsynaptic currents have a very specific shape. This shape can modeled by multiplying two exponentials of opposite direction together (subtract one exponential from 1). This shape allows PSCs to act as coincidence detectors. The sharp rise allows PSCs to be temporally precise. The long decay allows PSCs to overlap in time and summate to drive an action potential. The shorter the decay the shorter the time window PSCs have to summate. Different cells and different PSC types have different rise rates and decay rates. PV interneurons have very “fast” excitatory currents which means they need very tightly synchronized synaptic input. Pyramidal cell have “slower” excitatory currents which means they can summation input over a longer period of time. Inhibitory currents (at least on pyramidal cells) tend to be very slow.