Oscillation (?) following AP evoked by current clamp

The basics of how to develop, test, and use models.
Post Reply
kathryn

Oscillation (?) following AP evoked by current clamp

Post by kathryn »

I've noticed what seems to be a kind of oscillation following a single action potential, evoked by a long (1000 ms) depolarising current pulse. This behaviour appears on the voltage trace whether I build a 'single compartment' or use a very basic model cell (e.g. soma 20um x 20 um, hh; with or without an axon). Default values (imposed by NEURON) were used for all biophysical properties. The oscillation seems to be slightly worse in the absence of an axon.

At the end of the current pulse, the voltage drops back towards resting membrane potential (-65mV) as expected. However the cell then hyperpolarises a little before returning to -65mV.

Since I've relied on default settings and mechanisms (IClamp), I'm not sure where these responses are coming from. How do I eliminate them?

A screenshot, with the two responses circled in red:
http://picasaweb.google.com/frogologist ... 6705771874
ted
Site Admin
Posts: 6289
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: Oscillation (?) following AP evoked by current clamp

Post by ted »

kathryn wrote:I've noticed what seems to be a kind of oscillation following a single action potential, evoked by a long (1000 ms) depolarising current pulse.
That's a nice use of Picasa to share that figure. Unfortunately, egregious spammers have forced us to block posting of images to this Forum.

Your observation is quite perceptive. I wonder how many NEURON users have "seen but never noticed" these phenomena, which can be elicited with a wide range of stimulus and run times--e.g. they occur with much shorter current pulses and simulation durations (e.g. 20 ms and 200 ms, respectively).
I'm not sure where these responses are coming from.
(Ah, the spirit of inquiry is starting to stir.)
How do I eliminate them?
(Ah, how quickly it subsides.) "Close your eyes."

When a properly constructed model generates a result that violates one's expectations, it is usually one's expectations that are incorrect. Or, to paraphrase a character in the musical Chicago, "which are you going to believe--what you already 'know,' or your lying eyes?"

The phenomena are a kind of afterpotential (or after-potential). Afterpotentials following spikes were first observed in the era of extracellular recording from axons 60+ years ago. This ancient heritage is responsible for the peculiar terminology that you may encounter--"negative" and "positive" after-potentials are actually depolarizations and hyperpolarizations, respectively.

A pubmed search for the term after-potential turns up 244 articles, the earliest dating to 1953, but if you examine one of the articles from the 1950s or 1960s you'll find citations of relevant papers from earlier decades. Among textbooks from the past 10 years or so, "The Physiology of Excitable Cells" by the late David J. AIdley mentions spike after-potentials, but it's an exception in this era of molecular biology, where a lot of basic phenomonology is forgotten or ignored.

One easy and informative read that you can get for free is
NARAHASHI T, YAMASAKI T.
Mechanism of the after-potential production in the giant axons of the cockroach.
J Physiol. 1960 Apr;151:75-88.
As those authors note, extracellular K accumulation (which shifts EK in a depolarizing direction) is responsible for the negative (depolarizing) after-potential in squid and cockroach axons.

But the Hodkin-Huxley model generates after-potentials even if ion accumulation is omitted. To discover the underlying mechanism, create a single compartment model, then:
0. open a Voltage axis graph
1. open three State axis graphs and use them to plot (individually) m_hh, h_hh, and n_hh
2. set tstop to 50 ms
3. set your IClamp to deliver a 0.1 nA x 0.1 ms stimulus that starts at 1 ms

Now run a simulation and examine the time course of these variables. Pay particular attention to the interations between v and the gating variables; zoom in on the latter if necessary. Notice the first slight overshoot of h that peaks near 18-20 ms; why does this occur, what will it do to v, and will its effect on v be immediate or slightly delayed? Also notice the slight overshoot of n at about 20-24 ms; why does this occur, what will it do to v, and will this effect be immediate or slightly delayed? Also look closely at the later time course of h and n, and also of m, which also fluctuates more quickly than either h or n in response to changes of v.

Something similar occurs after the end of a long sustained depolarizing pulse, by a similar mechanism. And in both cases, there are related oscillatory changes in the excitability of the axon. You can explore this for yourself by using two IClamps, one to deliver a brief suprathreshold "conditioning" stimulus, and the other to deliver a brief "test" stimulus. For different test stimulus latencies, determine the current threshold for eliciting a spike, and you will (re)discover the episodes of hyper- and hypoexcitability that follow a single action potential. This has an interesting implication for the operation of neurons in networks: excitable cells may respond preferentially to excitation at certain interstimulus intervals, and less well to stimulation at other frequencies.
Post Reply