Abnormal brain development can lead to structural changes in the neocortex and subsequent neurological deficits, including epilepsy. One such cortical abnormality is polymicrogyria, which is an increased number of cortical folds, and disruption of the underlying lamination. These structural and laminar irregularities can be mimicked in rat by lesioning the brain just after birth, while the cortical layers are still forming. As in humans, this brain abnormality in rats is associated with an increased susceptibility to seizures. Our goal is to understand the neurological mechanisms that contribute to this hyperexcitability. Our previous studies suggest that there are both connectional rearrangements, and alterations in intrinsic properties of particular neuronal subtypes. Excitatory inputs to the malformed region shift to innervate the adjacent cortex, so that this surrounding region receives both its normal input as well as the reorganized excitatory afferents. This change occurs prior to the onset of epileptiform activity, suggesting that there are additional factors that contribute to the inception of population hyperexcitability. When overall inhibitory input to the epileptiform region is measured, it appears unchanged. However, when inhibitory interneuron subtypes are examined separately, they can be shown to have modifications in opposing directions. Fast-spiking interneurons normally provide powerful inhibition that impedes horizontal spread of activity through cortex. These neurons appear to be decreased in number or function in the malformed brain. Low-threshold spiking (LTS) interneurons normally synapse simultaneously throughout cortical layers of a single column and are far less-powerful, playing a more modulatory role. Within the malformed brain, LTS cells are increased in effectiveness, which may help to synchronize cortical activity. This type of synchronization is the first necessary step in the production of epileptiform activity.
We are in the process of developing a computational model of cortical columns, including these specific neuronal subtypes. With such a model, we expect to test each individual abnormality to determine whether it alone might initiate epileptiform activity. To be tested are: increased numbers of excitatory afferents; decreased numbers of fast-spiking interneurons; increased numbers, firing capacity, and connectional strength of LTS interneurons (each separately). There are a number of issues in developing a computational model of cortex where the choices in methodology are not clear. For instance, what level of detail is necessary in the individual neurons. We have currently chosen to model individual neurons with only a few compartments, since the questions being asked are about the network. A second issue is whether we need to test many different ways of connecting the neurons. In our current computational model we have used as much information as is available in the literature, primarily from physiological studies involving paired recordings. Despite this, there are still many unknowns in terms of the connectivity, that therefore require guesses. We have tested our modeled cortical columns on known behavior, including sleep rhythms, and under conditions of inhibitory neurotransmitter blockade. These computational ‘experiments’ show that the model acts similar to the actual cortex during biological experiments. Our hope is that through collaboration with modelers and/or mathematicians, a valid model will be developed from which we can make predictions about the next most appropriate biological experiment. This is a useful goal, since for instance, should the model show that reducing LTS neuronal activity would be the most effective way to limit seizures, there may be drugs available that would selectively control this specific abnormality within the malformed brain.
Cortical Connections to Computation: Can Calculations Tell us which Biological Abnormalities Contribute to Epileptiform Activity?
|Back to Seminar Page