A study of Neural Circuitry in Microdeletion Syndrome

Abstract

Abstract
Introduction. The Brain is an amazingly complex piece of biological hardware, capable of massive information processing. It relies heavily on the delicate interactions between huge distributed networks to work properly. When these interactions are altered there can be any number of consequences. These carefully tuned networks are largely dependent on the internal interaction of multiple genes. When these are damaged, changes in function can be seen. For instance, the 15q-microdeletion syndrome has been implicated in the possible development of some disabling conditions, for instance schizophrenia. It has been hypothesized that a pathway between the hippocampal region and the prefrontal cortex (PFC) has altered connectivity when compared to the same pathway in non-15q-microdeletion brains.

Methods. To check this hypothesis, a line of transgenic Human-15q-microdeletion mice were optogenetically modified and stimulated while recordings were taken in their PFC and stored as neuron-files, which contain data pertaining to spikes, waveform and recording location. The data acquired from acute measurements from two groups (Wild-type and Transgenic) mice was used to construct spike-time datasets. These in turn were Z-score normalized to compensate for background noise.

Each neuron-unit was categorized as either an interneuron or pyramidal cell based on their firing waveform. The pyramidal cells were further divided as either excited or inhibited Pyramidal Cells. Following the classification, the datasets were classified as belonging to one of the three classes of cells. The number of each cell-class was counted within each animal genotype for population comparison.

Results. The neuron raster plots revealed significant divergences in the interneuron and excited pyramidal cell response to gamma-range (40Hz) Stimulus. The Inhibited pyramidal cells exhibited what may be a deeper depression in transgenic cells, though not significant. Theta-range (6Hz) stimulus responses had no significant differences, except for inhibited pyramidal cells, which suffered a greater firing rate depression. Additionally, the ratio of cells was different between the wild-type and transgenic populations, with the transgenic mice having a reduced number of interneurons.

Conclusion. Generally, the transgenic populations seem incapable of keeping up with the wild-type at gamma-range entrainment. This has been previously observed in the behavior of 15q13.3MD mice auditory neurons. Interestingly the transgenic individuals seem to have lower interneuron counts, which has been observed in some schizophrenic human patients. Lastly the KV3.1 receptor blocker seems to target transgenic neurons specifically, exhibiting no alteration of activity in wild-type neurons.

Popular Abstract

Popular Summary
The brain is composed of millions of small computational units assembled in vast networks. When these networks don’t function quite the way they should, the effects for the individual can be severe.

The diseases caused by these malfunctions include conditions like epilepsy and schizophrenia. These can be very hard to diagnose, and every harder to find a root cause for. The difficulty partly lies in the vast amount of information exchanged by the computational units, the neurons, present in the brain. This information is exchanged between neurons in the form of electric spikes where the frequency of spikes encodes information. Today it is only possible to access the information indirectly by recording the spikes when the neurons fire using an electronic probe. This further degrades the signal making it even harder to make out from the background electric noise present from the rest of the brain shooting of signals. However, shanking people’s brains with electrodes naturally has some ethical problems.

This has prompted scientists to develop genetically modified mice (among other animals!) that have very similar conditions to the human ones. These animals, then, should hopefully be similar to affected people in both relative neuron activity, and network appearance. This needs to be verified, something that yet again requires statistical algorithms to make certain. The difficulty is compounded when you realize that to compare the network architecture you need to identify not only where each area is, but also what it’s composed of.

Luckily, there are ways to deal with this. Enter software analysis! By using statistical algorithms and self-adapting apps within an analytical computer environment, such as MATLAB, the large amount of data recorded from a brain can be broken down into much easier-to-understand patterns, allowing us to understand what is actually going in the brain. The patterns can then be used to sort each recorded neuron into distinct categories based on the properties of their spikes, and furthermore to extract the average activity of the place where the electrode was placed.

Using this, two groups of mice were used to get a bunch of recordings. One group was genetically modified to show schizophrenia-like brain activity, and one group was normal. After running the recorded data through all the algorithms, the author of this thesis found that the modified mice had a noticeable difference in the sizes of two different types of neurons, something mirrored in counts taken of the brains of people with the analogous condition in humans.

The mice were then treated, to see if it was possible to make the modified mice brain activity more similar to the unmodified ones. It turns out that the treatment not only makes the modified brains work more alike to the unmodified one, but the treatment also has very little effect on the unmodified brain, meaning it’s very specific to the condition.

In conclusion, it turns out that the mice do in fact reflect the relative changes found in humans with this mutation, and that the treatment can be used tread schizophrenia in mice, and hopefully one day in people.

• author: Holmberg, David
• supervisor: Palmi Thor Thorbergsson; Pär Halje; Jonas Thelin
• organization: Department of Electrical and Information Technology
• course: EITM01 20181
• year: 2020
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering
• keywords: Signal Processing, Analytical Software, Large Data Set, Classification, Sorting, Neurotechnology
• report number: LU/LTH-EIT 2020-783
• language: English
• id: 9027137
• date added to LUP: 2020-08-26 10:23:07
• date last changed: 2020-09-10 03:40:50