Talk:Models of deep brain stimulation

In this article authors have presented several computational models of deep-brain stimulation (DBS). In general they have done a very good job in carefully selecting models to highlight different modelling approaches. There are two key things that I missed while reading the review 1. The authors have reported about different models but only occasionally they have evaluated the models in terms of model predictions and match with the experimental data and in terms of what the model could not explain. Similarly they did not highlight if the model actually translated in to any changed in the DBS protocols. 2. This kind of review is a nice opportunity to set out new challenges to the modelling community so I would like to encourage the authors to bring up the challenges of the next generation of models.

Overall the review lacks flow and the contents could be organised in slightly better way to improve the readability. I also noticed that some important references are missing. My suggestions are listed below arranged according to the sections in the article.

——

• • section: Parkinson’s disease – an introduction

- Para 1: Last line is out of context. A better place for this reference to the figure is in the next para when the three pathways are introduced.

- Para 2/3: I think the text needs to be organised in a better way. The second line gives a suggestion that the function of the three pathways will be described. But next lines describe the role of GPI.

- Para 2-Last line —> Typo. GPI neurons project to part of thalamus involved in competing motor patterns.

- Para 4: Right now it is a bit out of context.

- For the first subsection I think the contents need to be organised in logical manner. I suggest that first the authors describe PD in terms of behavioural changes, then in terms of altered chemistry of the BG and finally in terms of the circuitry of the BG.

• • section: Imbalance between basal ganglia pathways

- The description is qualitative. When authors talk about altered neuronal activity in the BG subnetwork it would help the reader to provide quantitative information about the change. - Authors have completely ignored any change in the activity of striatal neurons. Since at least two pathways of the BG start in the striatum it would be necessary to describe how the activity of striatum is changed [see Costa et al. 2006, Mallet et al. 2008]

- The beta band oscillations are the most prominent signature of the BG activity but very little has been described about it.

- The first needs to be supported with appropriate references. Also the word ‘initially’ is a strange choice — It would be better to indicate some time frame.

- The line “Furthermore, the inability to decorrelate motor sub-circuits may explain why PD patients have difficulty in performing two simultaneous movements.” needs a reference.

- Authors should cite Tachibana et al. 2011 regarding the issue of increased bursting in STN and GP regions.

• • section: DBS as an effective treatment for Parkinson’s disease

- The second para in the section should be put under a different section heading as in this para authors are discussing putative mechanisms of how DBS works and not how effective DBS is. Moreover, authors should mention the recent attempts to tease apart the DBS mechanisms by means of selective optogenetigic stimulation of different DBS targets [e.g. Gardinaru et al. 2009]

• • section: Computational models of DBS

- The opening line of the para does not connect well with the next one. The models of BG were never developed with a goal to understand the effect of DBS. The goal of PD models maybe partly to understand DBS.

• • section: system level models

- A better expression for this class of models will be a firing-rate based models or neutron-mass models. And such models can never say anything about the neuronal assemblies.

- Before addressing a specific example of systems-level model I think authors should describe what these models are. I think the model by Leblois et al. 2006 could be used for this purpose. - I think a figure describing the Moroney et al 2008 could make it easy for the reader to understand the model - Moroney et al. 2009 is not in the reference list

• • section: Detailed models

- I am not sure if the models that are presented in this section can be termed as ‘detailed’. Yes, the neuron models were endowed with ion channels but the neuron morphology was either not present or was very simplistic. Similarly these models usually consisted on few 10s of neurons.

• • section: Amplitude and frequency dependence

- Author should also discus if the model predictions about the DBS amplitude and frequency are correct.

• • section: Testing different target and novel stimulation paradigms

- I think this should be a separate section and not bundled together as a part of the detailed model. In fact the models described in the Biophysical models section are more detailed than the models described in this section. - I would like to point the authors to other novel stimulation paradigms described in Kumar et al. 2011 [open loop stimulation] and Vlachos et al. 2016 [closed-loop stimulation, similar to the Rosenbaum and Pikovsky model but addresses the problem of stochastic oscillations which are more similar to the beta oscillations.]

• • section: Towards smarter DBS

- Here authors should mention the recent event-driven stimulation paradigm [Rosin et al. and Little et al.]

- While the models have been discussed their limitations have not been spelled out.

- Overall the English needs to be improved — I suggest that a native speaker edits the text.