Schizophrenia Research Forum - A Catalyst for Creative Thinking
Home Profile Membership/Get Newsletter Log In Contact Us
 For Patients & Families
What's New
Recent Updates
SRF Papers
Current Papers
Search All Papers
Search Comments
News
Research News
Conference News
Plain English
Forums
Current Hypotheses
Idea Lab
Online Discussions
Virtual Conferences
Interviews
Resources
What We Know
SchizophreniaGene
Animal Models
Drugs in Trials
Research Tools
Grants
Jobs
Conferences
Journals
Community Calendar
General Information
Community
Member Directory
Researcher Profiles
Institutes and Labs
About the Site
Mission
History
SRF Team
Advisory Board
Support Us
How to Cite
Fan (E)Mail
The Schizophrenia Research Forum web site is sponsored by the Brain and Behavior Research Foundation and was created with funding from the U.S. National Institute of Mental Health.
Live Discussion Transcript


Posted 20 September 2008

E-mail discussion
Printable version

Live Discussion: New Approaches to Cognition in Mental Illness

Return to Discussion Text

Attendees/Participants

Deanna Barch, Washington University, St. Louis
Robert Bilder, Semel Institute, University of California, Los Angeles
Nick Cuneen
Carla Gallo, Universidad Peruana Cayetano Heredia, Peru
Sharon Graw, University of Colorado, Denver
Ruben Gur, University of Pennsylvania
Hakon Heimer, Schizophrenia Research Forum
Julia Heinrich
Danny Koren, University of Haifa
Todd Lencz, Zucker Hillside Hospital, Glen Oaks, New York
Fred Sabb, Semel Institute, University of California, Los Angeles
Nico Stanculescu, Schizophrenia Research Forum
Joseph Ventura, University of California, Los Angeles
James Walters, Cardiff University, United Kingdom
Victoria Wilcox, Schizophrenia Research Forum

Note: The transcript has been edited for clarity and accuracy.


Hakon Heimer
I would like to introduce and thank our chat leaders, Deanna Barch, Fred Sabb, and Bob Bilder, for bringing this discussion to SRF.

For the first half hour or so, we'll be discussing CNTRICS (Cognitive Neuroscience Treatment Research to Improve Cognition in Schizophrenia). In the informal spirit of this event, and because I think most people here are at least acquaintances, I will only mention that Deanna Barch is a professor of Psychology, Psychiatry and Radiology at Washington University, St. Louis, where she is principal investigator at the Conte Center. What is most pertinent is that, with Cam Carter, she is the Principal Investigator for CNTRICS. Both Fred Sabb and Bob Bilder are at the Semel Institute, UCLA.

I'll now turn the floor over to Deanna.

Deanna Barch
Hello, all; thanks for joining the chat. One thing we were hoping to get out of the chat today is ideas/suggestions/questions regarding the future of the CNTRICS initiative. This meshes with our move into discussions of Phenowiki, as this is a potentially valuable way in which to move on from the start of CNTRICS.

Hakon Heimer
Deanna, could you say in a sentence or three where CNTRICS stands on its timeline?

Deanna Barch
Okay, to follow up on Hakon’s question, CNTRICS was started about two years ago to help facilitate the translation and inclusion of constructs and measures from cognitive neuroscience in the process of developing and testing pro-cognitive drugs in schizophrenia.

We have now completed a series of three conferences. The first focused on identifying the constructs from cognitive neuroscience that had the most validity and relevance to schizophrenia. The second focused on identifying the psychometric and practical challenges that such measures face in the translation process. The third meeting focused on selection and prioritization of tasks. Thus, we now have a set of constructs and a set of high-priority potential measures of these constructs.

Robert Bilder
Regarding Deanna's original question (suggestions for the future in the post-CNTRICS era): one thing I wonder is whether mouse-to-man and monkey-to-man translation can benefit from laying out more information about the paradigms used in each species, how we believe they are supposed to translate and accumulate evidence either supporting or refuting their validity. CNTRICS so far has focused mostly on the human paradigms, right?

Deanna Barch
Yes, Bob, CNTRICS has mostly focused on humans.

Fred Sabb
At the same time, I was impressed with how much the animal people at the third CNTRICS meeting contributed.

Deanna Barch
However, we are working on "CNTRICS 2," and a major focus will be mouse-to-man (or person-to-mouse and person-to-monkey) translation, driven in large part by the enthusiasm of the animal folks.

Fred Sabb
I think also that people were very enthusiastic about bringing the CNTRICS message to a larger community.

Deanna Barch
I should also note, though, that much work remains on the human side in terms of the need for the actual translation process to occur.

Robert Bilder
Given the centrality of transgenic models, the mouse-to-man seems particularly crucial.

Deanna Barch
Yes, mouse-to-man is particularly relevant for pharmacology as well. There was one RFA, and potentially a program announcement in the future, but there is really a need for folks to take up the translation challenge empirically.

Hakon Heimer
Fred, is there a model in particular that you could mention by way of example?

Fred Sabb
Hakon, in terms of models, everyone mentioned the Web, of course.

Hakon Heimer
Ah, I see. I thought you meant that there is an animal model that has some traction or obvious value.

Robert Bilder
It also seems like a lot of effort is dedicated to finding tasks that mice can do like humans. Some of our workgroups in the Consortium (Consortium for Neuropsychiatric Phenomics) have wondered if this is backwards, i.e., it's usually easier to get humans to do stuff mice can do than the other way around!

Deanna Barch
Bob, it is true that finding ways to get humans to do mouse models is doable, but what about the construct validity? If what mice can do is not what may be impaired in schizophrenia, is it like looking for keys under the streetlight? This is always the tension for me.

Fred Sabb
I think there are some particularly interesting cognitive tests that translate very well from mouse to man.

Robert Bilder
Some of the mouse memory models might be helpful starting points.

Fred Sabb
To fully bridge the two conversations, part of CNTRICS 3 really talked about translation of tasks like “stop-signal.”

Victoria Wilcox
If mice can do it, would there be enough variability in humans to study it?

Deanna Barch
Yes, mouse episodic memory models have a nicely accumulated body of data that speaks to the homology of mouse-to-man models.

Robert Bilder
Also, I would think while some of the aspects of schizophrenia may be hopeless for rodent models (i.e., persecutory delusions), whatever results from, say, DISC1, dysbindin, or neuregulin variants is likely not a “high-level” symptom but rather a more basic dysfunction.

Hakon Heimer
Bob, mice have working memory? ::big grin

Robert Bilder
Yes, mice have working memory; it just doesn't work as well as humans (usually?).

Hakon Heimer
Bob, Deanna, okay. I wondered because of the debate the neuroanatomists frequently have about whether mice have a prefrontal cortex.

Deanna Barch
So, doesn't this raise interesting issues about how we know when we are actually measuring a homologous process in humans and animals?

Robert Bilder
Hakon, mice don't manipulate too well, but they do have maintenance (briefly).

Fred Sabb
I think the relevance of mouse models is also something that needs to be continually discussed—and how the phenotypes overlap across species.

Deanna Barch
This seems a critical issue to be addressed going forward, as a lack of homology may sometimes be the basis for failed mouse-to-man translation.

Robert Bilder
Deanna, yes, I think that honing in on precisely what is the homology is crucial.

Deanna Barch
This is one of the things we were going to try to address in a series of animal/human translation conferences—developing a consensus in the field as to what constitutes homology, how would you know what it looks like, and what would your criteria be?

Robert Bilder
Right. At what level we look for the phenotype makes a huge difference. The big virtue of the mouse is that we can look directly at cellular and signaling pathways, etc.

Fred Sabb
I think CNTRICS did a good job of chatting about some of the important criteria for cross-species.

Ruben Gur
While I would generally go along with the belief that nothing is uniquely human, those who dispute that point to regions (such as the prefrontal cortex) that barely exist in mice, if at all. Do mice have "theory of mind"?

Robert Bilder
Ruben, do you think there is a tractable molecular biology of the "theory of mind"?

Ruben Gur
Why not, Bob? I am convinced that dogs have theory of mind. I'm not sure about mice, but I know a mouse researcher, Ted Brodkin, who thinks they do, and we are thinking of ways to prove that. For example, would a submissive mouse try to hide food from a dominant mouse (as some monkeys can do, according to some recent work)?

Deanna Barch
There is actually a growing set of work in mouse/rat models on such interesting "building blocks" of social cognition that might have a more tractable molecular biology. (I agree that theory of mind probably does not!)

Fred Sabb
In talks we've had amongst people at UCLA, it seems like it’s an ongoing discussion as both fields progress.

Robert Bilder
I guess a big issue is the extent to which we want the “behavior” to be homologous, or something more basic, like the behavior of neurons and neural systems.

Deanna Barch
Bob, interesting point! In other words, if we are interested in glutamate functions, is the level of homology at whatever process glutamate influences in the mouse, etc.?

Robert Bilder
Right, like some of Ty Cannon's work on DISC1 mouse (Li et al., 2007; see SRF related news story). There is behavioral data on working memory but also linking data from cell physiology.

Fred Sabb
I think that complex phenotypic models that span multiple levels are essential.

Deanna Barch
I think this discussion points to the importance of modeling at multiple levels, to help get a handle on where the homology is—what Fred said!

James Walters
I think the question of homology is crucial. A task can be formally equivalent in a neuropsychological sense without necessarily tapping equivalent functional systems, but arguments can be made for both being desirable criteria for homology. Would equivalent lesion mouse models and functional imaging data in humans be required?

Deanna Barch
James, good points. It seems like the homology criteria have to be at many levels, or perhaps criteria for different types of homology.

Robert Bilder
Ah, yes, Ruben, I'm confident we can find stuff that mice do that is particularly human, but I wonder if we looked at it from their (the mouse's) perspective, we would see it the same way.

Ruben Gur
Chances are that if we could ask the mouse, we won't because we'll see that as "introspective garbage"; isn't this what we do with humans?

Fred Sabb
Well, I think this is an excellent transition into how the Phenowiki can help.

Deanna Barch
Fred, tell us how!

Fred Sabb
It will provide a framework for how to have these types of discussions online. We really started to work on Phenowiki after starting to look through the available information on heritability and psychometric properties for the construct “cognitive control.”

Deanna Barch
Won't it do even more than that, though, in terms of linking multiple levels of analysis?

Fred Sabb
And as it exists now, it can support chatting about complex phenotypes and interaction between researchers from different fields.

Robert Bilder
I am actually sitting here with Stott Parker who has an R01 entitled "the Hypothesis Web.” Part of his idea is that multilevel modeling (i.e., from gene to protein to cellular systems and on up) is critical to establish correspondences.

Deanna Barch
Isn't that, in part, what we are discussing, the fact that there are many levels of analysis to understanding phenotypic manifestations of cognitive and neural systems?

Robert Bilder
Right, Deanna, and one classic limitation, I think, is that we tend to do experiments one pair of variables at a time, while convergence may require showing homologies across many levels. Otherwise, we have no clue as to whether the similarities are only superficial.

Fred Sabb
Deanna/Bob, yes, the Phenowiki will interact with Stott's Hypothesis Web.

Deanna Barch
Bob, so how does behavior fit into such a web—helpful but not necessary?

Fred Sabb
Deanna, we think behavior is an essential level, but not the only one.

Robert Bilder
Since we're all interested in behavior, that should be included always. :)

Deanna Barch
Bob, I can't find the right words to articulate, but given that we have greater access to one level of analysis in one species (e.g., model), and perhaps greater access to another level of analysis in another species (complex behavior in humans), how do we decide whether we have useful homologies when one or more levels don't quite match up?

Robert Bilder
Deanna, I'm optimistic that we can find the right behavioral parallels, particularly if we constrain these by lower levels.

Ruben Gur
One problem with mouse researchers is that they shy away from testing "batteries." They are very worried about altering the animal by each test. I've been trying to push folks here to do a "panel" of tests, and they are squirmish about it. There have to be some changes in their world (including assimilation of multivariate statistics) before they will be on board with us human researchers.

Robert Bilder
Ruben, I wonder if some of the mouse phenomics folks, like Rob Williams currently at Tennessee, might be willing.

Deanna Barch
Ruben/Bob, our mouse folks haven't been so hesitant to do "batteries," but the problem is that they often want to do the same battery for many different questions!

Hakon Heimer
While we transition to Phenowiki, let me ask Deanna if she has some major point that she'd like to raise vis-à-vis CNTRICS. All, also, does anyone have any particular thoughts to leave behind on CNTRICS?

Deanna Barch
Hakon, I would find it useful to hear suggestions as to how to stimulate researchers to want to engage in the translation of the types of cognitive neuroscience tasks nominated at the final CNTRICS meeting.

Fred Sabb
Deanna, plus, if we start to work out some of the quantitative relationships between levels, we can be more confident in our models. I also think that bringing some of the CNTRICS findings to the Phenowiki will allow others outside the CNTRICS group to have an important voice in the future directions.

Robert Bilder
Deanna, regarding stimulating research that can follow on CNTRICS, I wonder if you all plan to set up a repository so other researchers can download or access how to do those tasks?

Deanna Barch
Fred, do you think we need to be doing something more than the input process on which we are currently working?

Fred Sabb
Deanna, I think Bob just hit on it a bit.

Hakon Heimer
Fred, would you like to reiterate a couple of things that you hope to communicate/ask in this session?

Fred Sabb
Hakon, I think critical to this current discussion is bridging CNTRICS and Phenowiki, and how to get more people involved.

Victoria Wilcox
Deanna, do you see barriers to getting researchers to engage in the kind of translational work you mention? If so, what are they?

Deanna Barch
Victoria, well, I see a couple of barriers: 1) funding (always!), 2) getting basic scientists hooked up with clinical scientists in meaningful ways (some of that happened in CNTRICS, but more is better).

Robert Bilder
Deanna, is that RFA you mentioned still alive? It seems that NIMH remains quite interested; I wonder if interagency—across various NIH institutes and centers—there might be even more enthusiasm. This seems to be on the Neuroscience Blueprint, no?

Fred Sabb
Deanna, I think if besides support on the effect input side, development on the “wiki” side in the current wiki environment will allow people to investigate interactions between different phenotypic constructs.

Deanna Barch
Bob, part of the issue is that the tasks are essentially not ready for that, in that there currently are not "common" or "accepted" forms in the way that there are for most neuropsychological tasks. Well, I should moderate that; there are versions that could be available, maybe as a starting point for modification? I think most of the currently available versions are not optimized (that is the point of the RFA), so one would want to frame them as available starting points.

Hakon Heimer
All, can I ask if anyone here has explored Phenowiki, and if they could offer a researcher's "beta" view?

Fred Sabb
Phenowiki was initially conceptualized as a place where researchers could examine phenotypes to choose the best one for their particular needs.

Robert Bilder
I would think many researchers might have their own unique visions of customization/optimization. If there were, say, a freeware "mall" of scripts, some investigators might get over an energy hurdle faster (and overcome the inertia of using what is in their filing cabinets!).

Fred Sabb
This could be a great place for CNTRICS tasks to grow.

Deanna Barch
Bob/Fred, great suggestion. Maybe we can couple that with the "how-to" manuscripts that are going to be part of the write-up of the third CNTRICS conference.

Fred Sabb
Deanna, yes, I think that would be spectacular, because it would also give a larger community a voice.

Robert Bilder
Deanna, yes, I think there were a lot of great ideas in CNTRICS that go beyond the “beaten path” of much schizophrenia research on cognition, so making the methods easy to access might stimulate folks to incorporate them into their own research. And Fred, they could then use Phenowiki to see where the evidence exists that supports validity (or threatens validity) and determine what next experiment might really be a blockbuster!

Fred Sabb
Bob, yes, that would be excellent—using Phenowiki to better point out the weaknesses in current models/theories.

Deanna Barch
Bob/Fred, is there a way in the Phenowiki to link things more at the level of construct rather than at the level of task?

Robert Bilder
Deanna, yes, one of the key components of Phenowiki is that it represents studies at the task variable, task, and concept level.

Fred Sabb
Deanna, currently that’s in the wiki side.

Ruben Gur
Deanna makes a great point. My problem with many of the current computerized tests is that they basically copy the old ones and do not add a stronger link to basic neuroscience. There should be more editing going into the Phenowiki, or at least some effort at classification.

Fred Sabb
One additional plug for a recently funded project by Dr. Poldrack to establish a “cognitive atlas”; this will encourage the types of interaction both Ruben and Deanna are talking about.

Robert Bilder
Fred/Deanna/Ruben, in theory, Phenowiki can support exactly the kind of validity testing across species that would help us determine whether there are convergent threads establishing homologies.

Fred Sabb
Bob, yes, we hope that eventually Phenowiki can be a gateway to supporting that type of testing.

Robert Bilder
Or as Fred points out, the "next generation" elaboration of Phenowiki should be part of The Cognitive Atlas project (R01 with Russ Poldrack as PI).

Deanna Barch
Bob, yes indeed. In fact, could one think about (as an example) "operationalizing" some major theories that try to do that in the Phenowiki as a way of illustrating how that would work?

Fred Sabb
Bob, and yes, the testing may likely come from the Stott Parker “Hypothesis Web.”

Ruben Gur
Of course, any of that, if we are to do it seriously, will require some funding, no?

Fred Sabb
Ruben, yes, funding is obviously important, but buy-in by the community is also extremely important...hence this discussion. :)

Robert Bilder
Deanna, it would be great to have us focus on some initial "proof of concept" behavioral mechanisms and related procedures.

Fred Sabb
Bob, yes, we expect that Cognitive Atlas will be the next generation of collaborative editing for Phenowiki.

Deanna Barch
Bob/Fred/Ruben, I think a few clear examples would help the buy-in idea. Those of us who are newer to this technology sometimes need some concrete examples to illustrate how it works and to stimulate thinking and application to one's own questions.

Fred Sabb
Phenowiki will be the input gateway for quantitative effects.

Robert Bilder
Ruben, with Parker's Hypothesis Web and Poldrack's Cognitive Atlas, there are at least two R01s in place, but clearly the success of such ventures demands community engagement as Fred points out. Then it would be great, too, to establish larger efforts that could enable collaborative knowledge aggregation.

Hakon Heimer
Bob, Fred, it seems like for maximum effectiveness at the subclinical phenotype level, you will want buy-in from many disease research groups—schizophrenia, MDD (major depressive disorder), BP (bipolar disorder), etc. Is that an issue that you think about, or is it the case that the wiki will go where the wikipedians want to go, and that's okay?

Robert Bilder
Hakon, that is right on! In our Consortium for Neuropsychiatric Phenomics, we had a "Cross Disorder Phenotyping" committee! The idea is that there are so many likely shared mechanisms that it may be most fruitful specifically to focus on those phenotypes that are not unique to any specific disorder.

Fred Sabb
Deanna, that’s a good point. We've been focused more on the software development (yuck) side. But good cross-level examples are necessary to push people along.

Robert Bilder
Deanna, I wonder if "cognitive control" could serve as an example.

Deanna Barch
Ah, a construct near and dear to my heart! It is certainly one with a nice body of data.

Fred Sabb
Bob/Deanna, yes, cognitive control is always at the center of my thoughts, so all the content has revolved around that to begin with.

Robert Bilder
Fred is being modest. His Molecular Psychiatry target paper focused on cognitive control.

Deanna Barch
Fred/Bob, oh yes, that was a good example of an example construct.

Robert Bilder
So with cognitive control, Fred's paper suggested that it had so far been assessed largely via the same measures previously attributed to other constructs (like working memory, response inhibition, response selection, and task/set switching).

Hakon Heimer
Bob, if only we had MDD (major depressive disorder) and BP (bipolar disease) fora where we could also discuss this!

Robert Bilder
Hakon, I'm confident that you could rebrand as "Psychiatry Research Forum" and get the audience!

Fred Sabb
Bob, yes, the Molecular Psychiatry paper would be a good starting point. It represents some aspects of our database development as well as a starting point for examining the construct at a latent level (through literature association).

Robert Bilder
Deanna, we have requested a supplement to support a "Mouse-Man Translation in the GWAS Era" conference. [Editor’s note: “GWAS” refers to genomewide association studies.] I guess we'll know soon if this is going to be supported.

Deanna Barch
Bob, focusing on high-throughput paradigms for GWAS? Bob, or broader, just framed in terms of what GWAS opens up?

Robert Bilder
I think that high-throughput phenotyping is a longer-term goal; the first steps involve gaining further consensus on what criteria will comprise adequate homologies.

Fred Sabb
Bob/Deanna, yes, when we talk with our mouse folks, we find that it takes some discussion before we start to agree on a common language. Then we can start homology. CNTRICS, I think, experienced this, too.

Robert Bilder
The reason to focus on GWAS is that we will soon be deluged with genetic variants that no one ever thought about before. And the issues are different than those that have been focused on, e.g., in neuropsychopharmacology.

Hakon Heimer
All, do you feel that GWAS adds special pressure to develop standard tests? For instance, once strong candidate genes are identified, mice are generated, and people need to get their studies out so they can get their papers into Science/Nature/PNAS, they will grab some extant testing paradigms and use them. It seems that this could produce some potentially wasteful research and also "codify" certain paradigms just because that’s what someone did first (to get that paper into PNAS before their competitors).

Todd Lencz
I think, given the rapidity at which genomewide association studies are progressing, and the lack of convergence towards a limited number of potential targets, a few simple, high-throughput consensus tests might be advantageous in the short term. Also in mice, the ENU (N-ethyl-N-Nitrosoureanitrosourea) mutagenesis project provides an opportunity for broad genomic screening but requires very rapid and scalable phenotyping.

Ruben Gur
Another adversary to tackle is the large group of gene smashers who don't think they need these "endophenotypes" for genomewide association. They say that the only real hurdle in our field from the genomics angle is power: "Just give me 50,000 people with schizophrenia and 50,000 controls, and we'll solve all your problems."

Robert Bilder
Ruben, that is a huge issue. I believe we will have to face a solid five years of disappointment from gene discovery (or lack thereof) for the more complex phenotypes before people agree more that we should have been drilling down to lower-level and multilevel phenotypes.

Fred Sabb
Ruben, hopefully as we better refine our phenotypes of interest, we will be able to show geneticists the importance of that, as Freimer (Bearden and Freimer, 2006; Freimer and Sabatti, 2003) suggested a while back.

Todd Lencz
I agree with Hakon. We must have short-term plans in place to keep pace with the surrounding technologies.

Fred Sabb
Todd/Hakon, yes, I think that’s an excellent point, and I think we're probably already far behind in development of good phenotypes.

Robert Bilder
Todd, glad we could draw you onto your keyboard.:) This perhaps brings us back to how we might best focus on some domains of cognition most ripe for developing rodent models. Here at the CNP (Consortium for Neuropsychiatric Phenomics), we are biased towards "memory mechanisms" and "response inhibition" mechanisms. Perhaps this does overlap well with some CNTRICS constructs?

Deanna Barch
Bob, yes, it clearly does overlap with memory, executive control, early perceptual functions, and reinforcement learning. There are actually a host of cognitive processes that could be very ripe for translation and that might engage some novel paradigms.

Fred Sabb
Even with the best intentions, it’s too easy to just pick poor but available paradigms.

Deanna Barch
Fred, poor but available was what got CNTRICS started for humans—can't keep waiting to play catch-up but need to get good paradigms into shape now! The same is true for animal models, but it means some willingness to put the cart before the horse or to be proactive.

Fred Sabb
Deanna, yes, I agree, catch-up is bad. We have to gain some traction to get ahead of people, which I think will require some of these informatics approaches.

Deanna Barch
Bob/Fred, what about additional forums for getting the word out about Phenowiki—animal conferences, pharmacology, etc.?

Fred Sabb
Deanna, yes, we need to do a better job of advertising. I’m hoping that with some CNTRICS backing, people will come in droves!

Robert Bilder
Hakon, even if we do not get the NIH Roadmap support, perhaps we can cobble something together at the December 2008 ACNP (American College of Neuropsychopharmacology) meeting, like a satellite?

Deanna Barch
Bob and Hakon, one thing to think about is going to places where folks might not be thinking as much about these things. ACNP has folks thinking at least some about it (though maybe not the Phenowiki part), but what about more basic science meetings, getting them engaged more in application to clinical questions?

Fred Sabb
Deanna, that’s a good point, too. We should not always be “preaching to the choir.”

Robert Bilder
Deanna, I wonder what fora might be good for that?

Deanna Barch
Bob/Fred, I think SfN (Society for Neuroscience) is a bit big, but that is one idea. What about smaller behavioral pharmacology meetings?

Hakon Heimer
Deanna, I think that the size of SfN need not be a deterrent. With the proper advertising, you can find a lot of rodent psychobiologists there. (I was once one!)

Deanna Barch
Hakon, SfN does seem a logical place where a lot of the relevant people go.

Robert Bilder
Deanna/Hakon, yes, SfN has almost everyone!

Fred Sabb
I feel like Phenowiki did make an appearance at SfN last year—I presented that darn thing everywhere. :)

Ruben Gur
I (and postdocs here) think that Phenowiki and PubBrain rock!

Fred Sabb
Ruben, thanks! We're really hoping that people can see the grand vision and want to contribute.

Hakon Heimer
Fred, I'd like to suggest a more hands-on approach, i.e., I get the sense from the non-response to my question that not many people here had a chance to test drive Phenowiki. Perhaps an actual tutorial at SfN….

Fred Sabb
Hakon, yes, I think that would be good.

Deanna Barch
Hakon, a tutorial and perhaps a couple of really concrete examples? All, I must go now; I have students banging on my door!

Ruben Gur
Deanna, make sure you have heavy-duty hinges!

Hakon Heimer
Fred or Nico has access to WebEx software that will allow you to walk through things if you have the energy to do it online. We could schedule a follow-up in the fall where you do this by WebEx.

Fred Sabb
Hakon, I think that would be even better. With SfN being so large, it’s easy for one thing to get lost.

Hakon Heimer
Fred, maybe you should also give away free pens and occasional t-shirts to Phenowiki users. Don't laugh; as Nico will tell you, it works!

Nico Stanculescu
Hahaha! It does!

Fred Sabb
Yes, it does pay to advertise; everyone get their grad students wearing them. We can have user stats on them for the heaviest contributors.

Robert Bilder
I have some PubBrain mugs still in my office!

Hakon Heimer
For those who need to go to lunch/dinner/back to bed, let's think about some sweeping last statements.

Deanna Barch
Hakon, I am one of those who has to go, but I think that we need continued ways of developing engagement of the basic scientists as well as the clinical scientists and a proactive focus on task development for the future rather than the past!

Fred Sabb
Thanks, Deanna!

Robert Bilder
To keep the momentum going from CNTRICS, it seems that the collaborative knowledge-building space can be beneficial, but we should really elaborate, as Deanna has said repeatedly, some well-worked-out examples.

Todd Lencz
Also, it might be helpful to think about what kinds of endophenotypes might have more readily constrained genetic architecture—e.g., GWAS of height demonstrate an extremely normally distributed phenotype that (not surprisingly) is under the influence of an enormous number of genes of very small effect. What cognitive subcomponents are most “unlike” g or height, and might (speculatively) be under the influence of a fewer number of genes of larger effect? Or perhaps this can only be identified at the circuitry level?

Hakon Heimer
Any thoughts on Todd's question—cognitive subcomponents unlike g or height? Or will I have to corral Todd into leading our next discussion on that topic? ::big grin

Ruben Gur
Todd makes an excellent point, but it's sometimes hard to know ahead of time which traits are like height and which are like eye color. Genetic correlations could help, when the sample size is right.

Robert Bilder
Todd/Hakon, I think if we do use multilevel phenotypes, where the high-level behaviors (say, some extinction processes) are tied to specific neural circuitry where we can also do the relevant physiology, then we can have a bit more phenotypic purity (acknowledging that the high-level behavior has lots of contributors). I think it is arguable that hardly any behavior will really have a "simpler genetic architecture," even though the expectation that endophenotypes will have a simpler genetic architecture than the “disease” phenotypes has long driven some endophenotype research, so we need to make the ties to mechanisms more consistently.

James Walters
I think that the CNTRICS approach would be particularly helpful for psychiatric genetics in that it could provide more hypothesis-driven genetic research, providing that the neuroanatomical and neurochemical pathways of the CNTRICS paradigms are known.

Fred Sabb
Todd/Bob, I think if we lay out some of the interactions between phenotypes at different levels, with quantitative data, we can select common pathways that could potentially lead to smaller gene contributions. And yes, CNTRICS does think heavily about imaging.

Hakon Heimer
All, is it silly to think in terms of, Can you throw imaging into the mix? Can you get enough people to PET image mice during cognitive tasks?

Fred Sabb
Hakon, I think that imaging has an important role to play across species, but until the tasks are better, it won’t maximize its potential. And yes, CNTRICS does think heavily about imaging.

Robert Bilder
Hakon, yes, the imaging piece is already a part of it. The PET stuff is perhaps even easier (given that we can specify more clearly phenotypes like receptor binding) than, say, fMRI where we are less sure what the “neural system activation” actually means.

Ruben Gur
PET in mice can be cumbersome, and autoradiography is a one-shot-mouse-dead deal, but fMRI in mice can be done at high field. Several recent studies even showed nice amygdala activation to an intruder paradigm.

James Walters
I think that the CNTRICS approach would be particularly helpful for psychiatric genetics in that it could provide more hypothesis-driven genetic research, providing that the neuroanatomical and neurochemical pathways of the CNTRICS paradigms are known.

Robert Bilder
James, right on. I think that is the great promise of trying to specify the actual mechanisms that we believe underlie these complex cognitive neurological procedures.

Fred Sabb
This gets back to complex phenotypes, where we need to understand things at multiple levels—not just imaging, but the pathways below and the behavior above.

Robert Bilder
Ruben is right, of course; I was thinking actually of the one-shot paradigm for mice (sorry, mice), but at least that can yield some definitive results! And in the mouse, the cellular physiology and, of course, expression data are a lot easier to obtain.

Todd Lencz
I think the Meyer-Lindenberg DARPP-32 paper (Meyer-Lindenberg et al., 2007) [Editor’s note: SRF related news story] provides a good example of very targeted set of hypotheses, circuits, etc. Specific engagement of subcortical structures might be easier to get a handle on rather than more diffuse cortico-cortical networks.

Robert Bilder
Todd, great example. We also have a group working now on GPR6 (see Lobo et al., 2007). There is specific expression in medium spiny neurons in the basal ganglia, and this gene seems important specifically for some response inhibition functions.

Hakon Heimer
Todd, that model was used for AKT1 in their most recent paper in the Journal of Clinical Investigation. Here's my little blurb that links to that.

Robert Bilder
For the CNP, we have been using a simple "7 Level" hierarchy (Gene, Protein, Cellular System/Signaling Pathway, Neural System, Cognitive Phenotype, Symptom, Syndrome). I think it can be proved from this, as you make the traversal from one level to the next, that there is little hope of landing on some simple genetic link to complex behavior. Even from gene to protein, the variance shared across levels is not so high! Finding the right path through this maze is, I think, the grand challenge and why we need informatics representations to help us sort the most relevant (and irrelevant) paths.

Hakon Heimer
All, a couple of years ago, I heard that there was someone reporting success in introducing transgenes to rats. Does anyone know any more about this (it was an Alzheimer’s disease gene like APP, I think)? That would seem to open up the field as far as drawing on animal models of cognitive paradigms.

Robert Bilder
Hakon, I don't know much about the transgenic rat models, but I think transgenic mouse creation is so well developed at this point that we really can benefit and find traction there. We continue to use rats for neurophysiology and neuropsychopharmacology studies where mouse behavior is just not well enough developed.

Hakon Heimer
Bob, yes, those rats have bigger brains. Are they smarter than mice?

Robert Bilder
Hakon, while I have not given them IQ tests myself, our colleague David Jentsch assures me that the rats are a lot smarter!

Hakon Heimer
Bob, that's because they have a prefrontal cortex.

Todd Lencz
The ENU mutagenesis project(s) that I mentioned earlier is seeking to generate a comprehensive line of mutants. It is an incredibly ambitious goal. As far as I am aware, though, they only have a few simple behavioral screens that might be relevant for schizophrenia.

Robert Bilder
Todd, yes, this seems to be the challenge of many "bottom up" driven projects; they leverage the new high-throughput capacities at genomic levels, but then we are left with little knowledge at the phenotypic level. I think that is the big push of "phenomics" research, hopefully fostering more attention on the phenotypes.

Robert Bilder
Hakon, and a lot more cortex altogether!

Hakon Heimer
But maybe the amygdala is really in charge! ::wink. One of my amygdalar friends, if memory serves, says that it has had the greatest expansion after neocortex, in primates, relative to other brain areas.

Robert Bilder
Hakon, while that may have been said in humor, I think there is a critical point, and this links back to one of Todd's points, that we may fare better by focusing on some clearer circuitry. The prefrontal cortex is a big place with a lot of systems, and despite it being an attractive target for cognitive neuroscience research over the last few decades, not so much of that research has focused so specifically on specific PFC subcircuits, which may be required to generate valid connections to lower levels of biology. Amygdala and other subcortical systems might be really great targets; indeed, this helps bring home the point that target phenotypes need not be "top-down" selections (i.e., "disease" or "syndrome" relevance may actually be interfering with progress). We might make more rapid progress if we focus on, say, the molecular biology of central amygdaloid nucleus function.

Todd Lencz
Relatedly, I find that the Allen Brain Atlas is mostly useful for determining hippocampus/not hippocampus or basal ganglia/not basal ganglia. Cortical expression tends to look pretty diffuse, at least to my untrained eye. Bob, how would you envision a genetic pheno-atlas relative to the Allen approach?

James Walters
I have to go now, but thanks for letting me sit in on a very interesting discussion. I’ll look out for developments! Hakon, amygdala and other subcortical systems might be really great targets; indeed, this helps bring home the point that target phenotypes need not be "top-down" selections (i.e., "disease" or "syndrome" relevance may actually be interfering with progress). We might make more rapid progress if we focus on, say, the molecular biology of central amygdaloid nucleus function.

Robert Bilder
Todd, I think the pheno-atlas is just like the Allen and others, but not constrained to three-dimensional space. Stott Parker, working on PubAtlas and the Hypothesis Web, however, has given a lot of thought to how we can "align" concept spaces and "co-register" the "maps" that actually reflect different concept spaces.

Fred Sabb
Yes, again, I think that relying on any particular level to drive the investigation is bad. We need to understand the complexity across levels.

Todd Lencz
Actually, that is way too large a question, especially since I have to go in a minute. But I am struck by the issue of complementary timescales: rapid/high-throughput versus detailed/highly controlled. I don't think either is "better" than the other; we need to be working in both directions without expecting any grand synthesis.

Robert Bilder
Todd, I agree and was not trying to suggest limiting the bottom-up efforts. Just to get more attention to higher levels!

Hakon Heimer
Do we have any burning questions remaining?

Robert Bilder
I need more coffee.

Fred Sabb
Plenty for another day?

Hakon Heimer
As I said before, the chatroom never closes. Coffee for Bob!

Robert Bilder
Thanks a lot, Hakon, for offering this opportunity!

Fred Sabb
Yes, Hakon, this was excellent. Hopefully we can plan some webcast walk-through after the summer.

Hakon Heimer
And also, wine—a toast to Bob, Fred, and Deanna for leading this discussion!

Robert Bilder
Sayonara!

Joseph Ventura
Much food for thought; thank you, all!

Fred Sabb
Excellent; thanks again to all who joined.

Julia Heinrich
Thanks; this was great!

SRF News
SRF Comments
Text Size
Reset Text Size
Share/Bookmark
New Schizophrenia Fact Sheet for Patients and Families

Latest BBRF Research Breakthroughs

Visit our Facebook page and our Blog.

Support the Brain & Behavior Research Foundation Today.

Research Participants
Collaborators
Copyright © 2005- 2014 Schizophrenia Research Forum Privacy Policy Disclaimer Disclosure Copyright