linking back to brembs.net






My lab:
lab.png
In the 2009 call for their starting grants, the ERC announced that they would be funding innovative 'Frontier' research with higher-than-normal risk of not meting the proposed objectives:
In particular, proposals of an interdisciplinary nature which cross the boundaries between different fields of research, pioneering proposals addressing new and emerging fields of research or proposals introducing unconventional, innovative approaches and scientific inventions are encouraged, as long as the expected impact on science, scholarship or engineering is significant.
The ERC peer review evaluation panels will look carefully at these aspects, in the full understanding that such research has a high-gain/high-risk profile, i.e. if successful the payoffs will be very significant, but there is a higher-than-normal risk that such research projects may not achieve their objectives.
All this fitted our proposal exceptionally well, we thought.

We wanted to investigate what the default mode of the insect brain is. Currently, humans and other primates are the only species where this phenomenon has been studied in some detail. The function(s) of the default mode network (DMN) are unknown, but alterations of the DMN have been associated with a large array of mental disorders, such as schizophrenia, Alzheimer's, autism and many more. I have first heard about the DMN in 2006, but the term has been around since 2001. The DMN discoverer, Marcus Raichle, argues that "The brain is not primarily reflexive" and talks about a 'paradigm shift'. Clearly, research into intrinsic (i.e. not stimulus-related) brain activity is a 'new and emerging field' of 'an interdisciplinary nature crossing the boundaries' of neurobiology, behavioral biology and psychiatry. Given this importance for our understanding of general brain function, it is a staggering gap in our scientific arsenal that no model system exists in which we can perform a functional analysis of a DMN. All animals show spontaneous behavioral activity and every brain that has ever been probed shows intrinsic activity. Thus, there is no reason why other animals shouldn't have a DMN as well. We have plenty of experience with spontaneous behavior in flies and a recent graduate here has spent his PhD thesis on constructing a two-photon microscope, performing experiments with it and programming the evaluation algorithms. On the microscope market, there is exactly one microscope, the LaVision BioTech Trim Scope II, which would be capable of acquiring 3D brain activity data at a reasonable frame rate of 10-15Hz. If we express neuronal activity reporters in every neuron of the fly, in principle, we should be able to record data from flies that would be structurally very similar to the fMRI data of Marcus Raichle et al. in humans. If that approach isn't both 'unconventional and innovative', what is?

Interestingly, the three reviewers all agree with my assessment that our project has the potential to significantly broaden our understanding of general brain function and thus meets all the ERC starting grant criteria quoted above:
Reviewer 1:
The project is conceptually interesting and original
Reviewer 2:
The project extrapolates from the popular concept of a default mode network derived from human work to propose a broad theory of interaction between operant and classical synaptic learning, basing the arguments on evolutionary principles and biochemical differences. The proposal consists in moving from man to drosophila and using sophisticated optical scanning methods and task free compared to stimulated flies (the description of behavioural manipulation techniques is fascinating). This is true frontiers research
Reviewer 3:
Ground breaking nature of research:
Yes, understanding background or resting activity is important.
Potential impact:
Recording from all different parts of the brain simultaneously with single cell resolution would be wonderful.
However, the reviewers rightfully point towards technical difficulties:
Reviewer 1:
the PI seems too optimistic in areas where he has little expertise. [...] One wonders whether Drosophila is really the best model for this work, as other genetic model systems (C. elegans, zebrafish larvae) could be imaged much less invasively and with far less constraint in their locomotion. Mechanistic studies are mentioned as a goal, as is modelling. Neither of them is explained in detail, but in any event these are unrealistic goals at this point.
Reviewer 2:
This is true frontiers research but it is difficult to judge whether what is proposed is technically possible.
Reviewer 3:
Without knowing where to focus, using specific driver lines to express the genetically encoded indicator, it will be almost impossible.
Maybe with the exception of the very last comment (we explained that we will not focus on any specific brain area, but on the entire brain - that's what this particular microscope is good at - and for these reasons not use specific but general driver lines), these are all very valid comments and of course it is not a safe bet project. I also agree that I don't have the necessary expertise to do these experiments myself, obviously. Therefore, we explained in our grant that the experiments will be performed by someone who has 6 years of experience with this sort of technique. Thus, in principle, all the technical requirements are met to conduct this research. Granted, we don't have any pilot data, but if we already had these, it would be a no-risk proposal, which is exactly why I wrote the ERC grant - because they advertised it as a "High-Gain/High-Risk" instrument. If I had pilot data, I could apply to any funding agency, it would be a slam dunk. Heck, depending on how well it worked, such data could almost be a paper by themselves.

Without such pilot data, I would not have submitted this proposal to any regular funding agency, because it is understood that only conventional projects get funded there. However, the ERC is trying to brand itself as a not-so-regular funding agency. Apparently, technically feasible but not shown to actually work is already too high-risk for the ERC. I wonder what they would consider low-risk then? Maybe the ERC isn't so unusual after all and much more risk-averse than they would like to seem? Sure, our proposal is only one out of 2873 applications, but is there any data that shows that the ERC actually is funding any high-risk research?

In the light of my last post, I hasten to emphasize that all three reviewers were highly knowledgeable and by and large I agree with their overall assessment (even though there are specific points I would argue against). All three reviews were thorough, understood the kind of research we proposed and seemed to know what technical difficulties we might face. The only thing I wonder is whether the ERC is advertising one thing but actually funding another? Now, it could be that the panel assessment of my personal standing and performance was the basis for the rejection. However, there is nothing in the document that suggests this would be the case. On the contrary, the panel only refers to the project and not to me:
As seen from the individual reports, the project was perceived as conceptually interesting and original, but technically daunting. Also, the proposal seemed optimistic in areas where the applicant has little expertise: without specific driver lines targeting individual neuron populations, there seems to be little chance to record interpretable data.
Overall the panel considers the proposal of good quality, however at a relatively low position in the ranking order.
Ironically, the panel picked up the one single erroneous criticism of the project as an example for the technical difficulties. We state it clearly in our proposal: "We will use available genetic drivers to drive fluorophor expression in all neurons." Even if we needed specific lines (which we don't) there would be tens of thousands available to choose from.

Is there any High-Gain/High-Risk funder left anywhere? That microscope costs around 600k € with a 5-year warranty and I'd also need the salary for the postdoc (because yes, I don't have the technical expertise). The proposal has been rated as high-gain by three reviewers and the ERC panel already, now does anybody know where the high-risk funder would be?
Posted on Tuesday 20 April 2010 - 15:45:14 comment: 0
{TAGS}

Render time: 0.1070 sec, 0.0033 of that for queries.