# User talk:Jrm228/Final Term Paper

## Contents

## Student Comments

Hi James!

A few things:

1. It looks like the bifurcation analysis section of your results is lacking. Make sure to update this with pictures of the bifurcation and what it means to the hypothesis.

2. In the results, you talk about how the replication of his model had to be cut off at 20 mV to make all the spikes the same height, when in reality the model does always reach 30 mV if enough plotpoints are used. Thus, this isn't really grounds for declaring issues in the model equations since Mathematica integrators could run it properly.

--Jessica Nash (talk) 14:53, 30 April 2017 (EDT)

## Instructor Comments

- Term paper survey submitted with term paper?
- Yes.

### Introduction

- Significance of problem?
- Yes, but needs more depth and relevant literature citations.

- Statement of hypothesis?
- Yes.

- List of references?
- Total of 8, 5 used in the Introduction. Far too few for a scholarly and in depth introduction.

- Properly formatted references?
- Yes.

### Model Description

- State variables, parameters and inputs to the model clearly distinguished?
- Yes.

- Term-by-term description of model components?
- Yes.

- Have you described model assumptions?
- Yes.

- Description of equation simulation?
- Yes.

### Results

- Results described and compared to original paper?
- Yes, but you did not replicate even a small neural network as Izhikevich had done.

- Have you addressed the original biological hypothesis?
- Yes, but this could have been done better.

- Figures and legends to show results?
- Yes; but you should have worked harder to make your model neurons have the same time scale as those in the Izhikevich paper.

- Discrepancies relative to original model?
- Discussed, but the Guckenheimer bifurcation diagram is simply shown, not explained.

- Uploaded Mathematica file?
- Yes.

### Discussion

- How well does the model support the original hypothesis?
- Reasonably well discussed, but could have been more in depth. The missing piece, which you could have looked at, was the Izhikevich developed his model based on a dynamical analysis; at that more abstract level, his model does capture many features of more complex neural models.

- Support for hypothesis and assumptions from other data in the literature?
- This was a serious weakness of the Discussion. Almost no relevant literature is cited. 5

- Limitations of results?
- Well discussed, but should have focused more on the dynamical systems analysis that Guckenheimer did. You show his diagram, but do not explain or discuss it at all.

- Discrepancies and how they affect conclusions?
- Discussed, but could have been in more depth.

- Relationship to other work in the field?
- Another serious weakness. What other models are out there? Have others used Izhikevich’s model? What recent papers cite his work? None of this is discussed.

- Discussion of future work?
- Yes, but could be much stronger.

### Overall Term Paper Quality

- How well was the model replicated?
- Network not replicated; rest of model was replicated.

- Based on the term paper, how well did you understand the material?
- Understanding could be sharper.

- How well written is the term paper?
- Nicely written; clear and easy to read.

- How hard was the model extension that you did?
- Moderately difficult; but this has already been done by Izhikevich.

- How good was the extension?
- This could have been better described.

- Mathematica code clear and well annotated?
- Yes.

- Mathematica code generates figures when evaluated?
- Yes.

On the whole, this is a good term paper. It would have been much stronger if more technical literature had been read and understood and used to frame the Introduction and especially the Discussion more clearly. It was fine to show the Guckenheimer diagram, but it was not discussed at all, which is very problematic, especially since this was critical to the extension.

--Hillel Chiel (talk) 09:27, 12 May 2017 (EDT)