Control over stress induces plasticity of individual prefrontal cortical neurons: A conductance-based neural simulation

User Rating:  / 3
Least InterestingMost Interesting 










Juan A. Varela1*, Jungang Wang1, Andrew L. Varnell& Donald C. Cooper1


PDF version


*Collaborate with this Author, click the communication icon


Behavioral control over stressful stimuli induces resilience to future conditions when control is lacking. The medial prefrontal cortex(mPFC) is a critically important brain region required for plasticity of stress resilience. We found that control over stress induces plasticity of the intrinsic voltage-gated conductances of pyramidal neurons in the PFC. To gain insight into the underlying biophysical mechanisms of this plasticity we used the conductance- based neural simulation software tool, NEURON, to model the increase in membrane excitability associated with resilience to stress. A ball and stick multicompartment conductance-based model was used to realistically fit passive and active data traces from prototypical pyramidal neurons in neurons in rats with control over tail shock stress and those lacking control. The results indicate that the plasticity of membrane excitability associated with control over stress can be attributed to an increase in Na+ and Ca2+ T-type conductances and an increase in the leak conductance. Using simulated dendritic synaptic inputs we observed an increase in excitatory postsynaptic summation and amplification resulting in elevated action potential output. This realistic simulation suggests that control over stress enhances the output of the PFC and offers specific testable hypotheses to guide future electrophysiological mechanistic studies in animal models of resilience and vulnerability to stress.


Traumatic and stressful life events contribute to psychiatric conditions including posttraumatic stress disorder (PTSD), anxiety and depression, but individuals respond differently to stressors1. An individual’s sense of actual or perceived control over stressors is regarded as a potent coping mechanism that can lead to resilience from trauma. Behavioral control over electric shocks (escapable shock or ES group) prevents the physiological and psychological consequences of lack of control (inescapable shock or IS group)1. It has been demonstrated that the underlying mechanism requires lasting changes in neurotransmission within the medial prefrontal cortex (mPFC)2.

We have recent unpublished data indicating that pyramidal neurons in deep layers of the pre-limbic subregion of the mPFC from ES group rats undergo plasticity that increases neuronal excitability compared to IS or home caged (HC) control rats. ES neurons have faster membrane time constant (TC), larger action potential (AP) amplitude, faster AP rise rate and a larger post-spike after-depolarization (ADP). However, the measure of the underlying conductances involved in this plasticity requires precise and extensive quantification at the cellular/molecular level of ion channels present in mPFC pyramidal neurons or direct measure of the many currents associated with the changes in conductances between the groups. Instead, a simple solution is to build computer models in which the conductances can be modified. Therefore, we utilized the neural simulation tool NEURON to examine the relationships between various conductances to fit the measure physiological data, and determine the conductance changes between the different groups. Test other measures of excitability that we were unable to test experimentally and make further predictions based on the conductance differences between the groups.


To account for the difference between the ES and IS/HC groups we modeled representative traces from each group (Fig. 1). The differences are necessarily a reflection of changes in the biophysical properties of the neurons3.We developed a simple Ball and Stick model and fitted the data to elucidate the underlying conductances. First, we fitted the IS passive response to hyperpolarizing current pulses (Fig. 1). We fitted the following parameters: membrane capacitance (Cm), input resistance (RI) and the hyperpolarazing-activated cationic conductance (IH), then keeping the membrane capacitance and IH values constant we fitted the ES passive response (Fig. 1), the reduced input resistance of the ES trace compared to the IS trace accounts for the difference observed in TC between the IS and ES groups. 



Figure 1. Modeling ES and IS conditions. Data (black) was first fitted (red) by setting the passive parameters using a hyperpolaraizing current pulse (bottom) and then the active parameters using the action potential trace (top). Left, fit for the IS condition, right, fit for the ES condition. B. The only active parameters that needed to be changed to account for the IS/ES differences (AP amplitude, AP rise rate and ADP) were the Na (top) and the T-type Ca2+ (bottom) channel densities. Top, increasing the Na channel density increases the AP amplitude (black bars) and rise rate (grey bars) but decreases the ADP (white bars), the ES model (red) has 63% more Na channels compared to IS model (100%). Bottom, increasing the Ca channel density dramatically increases the ADP (white bars) while the AP amplitude (black bars) remains the same, the ES model (red) has 185% more T-type Ca2+channels than the IS model (100%).

To account for the differences in the active properties we fitted the IS AP data, using the fitted passive parameters (Cm and RI) and IH, divided the process into two parts; we fitted the rising part of the AP adjusting the following parameters: Na, K-delayed rectifier (Kdr) and the A-type K channel (KA) and then proceeded to fit the whole action potential (Fig. 1), fixing the previous parameters and adding the T-type Ca conductance (CaT) and the non-specific voltage-dependent cationic conductance (ICAN). Afterwards small manual parameter adjustments were performed to fine tune the fit. To fit the ES data we started with the IS parameters and only make changes in as fewer conductances as possible, this was accomplished by changing two parameters: Na and CaT. An increase of the Na channel density results in an increase of the AP amplitude and rise rate(Fig. 1). The increase if the CaT conductances greatly increase the ADP (Fig. 1), together (63% increase on the Na and 185% increase on the CaT) are sufficient to account for the differences between IS and ES. 

We decided to test if the excitability extended to the processing of synaptic inputs by the neurons as well. We studied “synaptic boosting”, the increase in amplitude and duration of the synaptic response as the resting potential of the neuron becomes more depolarized; we delivered the same simulated synaptic inputs (conductance and place in the dendrite) for both the IS and ES conditions at different membrane potentials (Fig. 2), and found that synaptic boosting occurs at lower resting potentials (i.e. more excitable) in ES compared to IS, from a 6% increase in synaptic are at -70 mV to a 136% increase at -61 mV (Fig. 2). When synaptic responses arrive at the soma at a fast enough frequency that if the previous synaptic response is not yet over the response “adds” to the previous one further depolarizing the neuron. Ten identical synaptic inputs were delivered in the dendrite at different frequencies we found that synaptic summation increased the total synaptic area from 9% at 20 Hz to 19% at 50 Hz in ES compared to IS and the total synaptic maximum amplitude by 1.2 mV at 20 Hz to 5.2 mV at 50 Hz (Fig. 2). Both, synaptic boosting and synaptic summation, result in a ES neuron that is active in response to stimulus that would not produce any activity in a neuron from the IS condition (Fig. 2).


Figure 2. EPSPs are functionally enhanced in ES compared to IS. Synaptic boosting. The same simulated synaptic input was delivered while changing the holding potential to more depolarizing values for both the ES (red) and IS (black) model. The model predicts a larger effect of the synaptic boosting in ES compared to IS. Synaptic summation. Trains of the same 10 synaptic input were delivered at different frequencies for both the ES (red) and IS (black) models at the same holding potential, synaptic inputs in ES summate more than those of IS, as seen on the difference traces (ES-IS) the effect is greater as frequency increases. At more depolarized potential (-63 mV) the summation and boosting add, resulting in greater excitability.





We have produced a model that can explain the main differences between the IS/HC and ES neuronal properties by only changing three parameters. The leak or RI, responsible for the shorter TC in the ES group, the Na conductance which accounts for the increased AP amplitude and faster AP rise rate in the ES group and the CaT conductance which explains the larger ADP in ES pyramidal neurons. Further, we can use this model to predict that because in the ES group any synaptic input will be enhanced by synaptic boosting and by temporal summation of the responses.


The conductance-based model Neuron ( and all neuron files utilized files can be found at (Neurocloud link). The following neuron mods (accession #s) were obtained from ModelDB4: 121259 (Na), 19696 (Na), 135839 (IH), 112546 (Kdr), 108459(KA), 123815 (CaT) and 12631 (ICAN).


Expanded References:

  1. Amat J, Paul E, Zarza C, Watkins LR, Maier SF. Previous experience with behavioral control over stress blocks the behavioral and dorsal raphe nucleus activating effects of later uncontrollable stress: role of the ventral medial prefrontal cortex. J Neurosci. 2006 Dec 20;26(51):13264-72.
  2. Baratta MV, Zarza CM, Gomez DM, Campeau S, Watkins LR, Maier SF. Selective activation of dorsal raphe nucleus-projecting neurons in the ventral medial prefrontal cortex by controllable stress. Eur J Neurosci. 2009 Sep;30(6):1111-6.
  3. Hines ML, Morse T, Migliore M, Carnevale NT, Shepherd GM. ModelDB: A Database to Support Computational Neuroscience. J Comput Neurosci. 2004 Jul-Aug;17(1):7-11.
  4. K. Sidiropoulou, E.D. Ozkan, M. Fowler, F.J. White, C. Philips, and D.C. Cooper (2006) Dopamine modulates an mGluR5-induced depolarization underlying prefrontal persistent activity, Nature Neuroscience, 2009, 12(2): 190-199.
  5. Southwick SM, Vythilingam M, Charney DS. The psychobiology of depression and resilience to stress: implications for prevention and treatment. Annu Rev Clin Psychol. 2005;1:255-91.

Download NEURON files here

Comment or Share