There is a pressing need for new ways of thinking about more effective and specific treatments for neurodegenerative diseases such as Alzheimer’s and Parkinson’s. One approach that holds great promise is to understand how damage to the molecular and cellular underpinnings of the brain impacts the many aspects of cognition—including memory, attention and cognitive flexibility—that are so profoundly affected in these diseases. But so far, the huge advances made in understanding the fundamental neurobiology of the brain have not translated to success in the clinic. One reason for this is that the effect of molecular and circuit level interventions on cognition can only be investigated in animal models, but assessment of cognition in animal models is rarely done in a way that is directly relevant to humans. This is likely a major contributor to the repeated failure of potential treatments for brain diseases in patients despite seemingly adequate and appropriate data in mouse models demonstrating that candidate drugs should work.
The Mouse Translational Research Accelerator Platform (MouseTRAP) directly addresses the urgent and critical challenge of improving clinical translation and patient outcomes. MouseTRAP is centred on a touchscreen-based system that we have developed that allows us to assess mice on tests of high-level cognition that are very similar or identical to those used in human patients. The automation and standardization inherent in this approach achieves more accurate, robust, and reproducible phenotyping of mice than conventional approaches. MouseTRAP pairs these tests with cutting-edge technologies to record or manipulate neuronal, glial or neurochemical activity, which makes it possible to match—millisecond by millisecond—what is happening in the brain with human-relevant cognitive performance. This can be done in healthy mice or in our extensive catalogue of next-generation disease models, making MouseTRAP a state-of-the-art platform for assessment of robust, reproducible and human-relevant cognitive outcomes in mouse models, for either fundamental discovery research or development of knowledge-based therapeutic interventions.
MouseTRAP is a distinctive platform that operates within the BrainsCAN Rodent Cognition Innovation Core and is available to researchers on a fee-for-service basis. Various levels of support are available to users at different rates: MouseTRAP technicians can conduct a full touchscreen-based cognitive assessment in a mouse model, users can be given training and access to equipment to run their own experiments, or specialized technical advice can be provided to users who wish to bring new techniques to their own labs. In all cases, our highly qualified core staff provide advice on techniques, experimental design, the interpretation of the experiment and presentation of the methods and results for publication. This specialized service uses a collection of mouse lines that can be accessed by users to model neurodegenerative or neuropsychiatric diseases or manipulate genetically-defined populations of cells.
A program of continuous development is built into MouseTRAP, which ensures that our tools remain cutting-edge. This work includes development of new tasks and validation of mouse disease model-task combinations (e.g., if we know that a specific mouse model relevant to Alzheimer’s disease is impaired on tests A, B and C then this model+task combination can be used to test therapeutic interventions)
Watch Associate Professor, Dept. of Philosophy at Western University, Jacqueline Sullivan describe the MouseTRAP platform in a recent talk titled “Touchscreens, Open Science, and the Epistemic Community.”
The technologies incorporated into MouseTRAP require an unusually broad range of expertise—spanning from molecular biology to neurochemistry to animal learning theory—along with expensive specialist equipment. These technical and financial barriers make it difficult, and often impossible, for scientists to adopt the full range of these techniques individually in their own labs. Offering MouseTRAP services to academic and industry scientists democratizes these tools by facilitating access both to our unique technical expertise and our world-leading, high-capacity touchscreen cognition facility.
Mouse behaviour core facilities usually focus on conventional behavioural assessment techniques (e.g., open field, elevated plus maze, fear conditioning, water maze). Although we have such tests available in the Rodent Cognition Research and Innovation Core, MouseTRAP focuses on touchscreen testing because it is more robust, more reproducible, and more relevant to humans. MouseTRAP runs on the largest collection of mouse touchscreens in the world, which enables high-throughput assessment of cognition in up to 500 mice per day on the more than developed by our team. We invented the technology and have over 25 years of experience in its development and use, providing unmatchable expertise.
The table below is a simplification of Table 1 published in Sullivan, Dumont, Memar et al. (2020) which shows a selection of rodent touchscreen tests currently available and the cognitive functions they assess. For more details, please see the original publication or visit . Note that the flexibility of experimental parameters of the touchscreen systems continue to inspire researchers to develop and refine cognitive tests, resulting in news tasks being developed continually!
|
Cognitive Construct |
Touchscreen Test Available |
|
1. Learning & memory |
A. Pairwaise visual discrimination learning |
|
|
B. Object-location paired associates learning |
|
|
C. Working Memory: i. Spatial: 1. Trial-unique nonmatching-to-location 2. Continuous trial-unique nonmatching-to-location 3. Delay match to position ii. Non-Spatial: 1. Nonspatial nonmatching-to-sample |
|
|
D. Location discrimination |
|
|
E. Spontaneous (novel) object recognition |
|
|
F. Visuomotor conditional learning |
|
|
G. Automated search task |
|
|
H. Heterogeneous long sequence task |
|
|
I. Transverse patterning J. Category learning K. Transitive Inference |
|
2. Attention |
A. 5-choice serial reaction time task |
|
|
B. Continuous performance test |
|
|
C. 5-choice continuous performance test D. Sustained attention task E. Spatial probability task F. Flanker task G. Posner cuing task |
|
3. Decision making |
A. Rodent gambling task B. Risky decision-making task C. Probability discounting D. Effort-related choice/decision E. Effort-related discounting G. Fixed ratio discounting |
|
4. Cognitive flexibility & response inhibition |
A. Reversal learning B. ID/ED Set-shifting C. Extinction learning D. Contextual rule switching E. The STABFLEX task |
|
5. Emotional cognition & responding to reinforcement |
A. Progressive ratio test of motivation/apathy B. Response to negative (& positive) feedback C. Autoshaping D. Cognitive judgment bias |
Neurons communicate with each other, and with astrocytes and microglia in the brain, via neurotransmitters including glutamate and GABA. Fine tuning of this communication depends on a number of other neurotransmitters or neuromodulators, including dopamine, acetylcholine , serotonin and noradrenaline. Although broad relationships between neuromodulator activity and cognition are accepted, we know very little about the precise roles and mechanisms by which these neuromodulators regulate different cognitive domains. In addition, recent observations have raised the possibility that neuromodulators can also signal to astrocytes and microglia and that glial cells can in turn signal back to neurons, but how this reciprocal signaling regulates cognition is unknown. These gaps contribute to our poor understanding of the mechanisms by which treatments fail to restore synaptic and neurochemical pathways to improve cognitive function in disease.
We have a pipeline of more than 20 novel fluorescent biosensors (many generated by our collaborator Yulong Li) that enable us to observe neurochemicals and/or cellular activity with fibre photometry in vivo while mice perform touchscreen-based cognitive tests. This innovative approach builds on a G-protein coupled receptor scaffold for a given neurotransmitter, which is mutated and fused to conformational sensitive fluorescent proteins. These sensors respond with large fluorescent changes on millisecond time scales—a step change in sensitivity and temporal resolution compared to traditional methods such as microdialysis—and enable us to pinpoint neurochemical or population-level neural activity underlying specific stages or aspects of cognition with exquisite sensitivity. They can be expressed in distinct brain regions and genetically-identical cell types using Cre mouse lines or promoters packed in viral vectors.
Automated touchscreen behavioural tasks in rodents enable systematic high-throughput and translational cognitive assessment with standardized outputs that facilitate data reproducibility, analysis, and dissemination. Traditionally, lesion studies, in vivo electrical stimulation or infusion of drugs affecting local cellular activity within the brain has been used to investigate circuitry and connectivity function underlying behaviour. Unfortunately, these approaches lack of either cellular or temporal resolution to match millisecond scale events correlated with the onset of environmental contingencies. Over the last decade, a generation of genetically encoded proteins that change conformation in the presence of light (opsins) have been developed to alter excitability of specific cell populations in discrete brain regions. The use of this experimental tool is known as ‘optogenetics’, and it has enabled investigators to address whether specific cell populations or brain regions underlie behaviors relevant to psychiatric and neurodegenerative disorders. Moreover, because the optogenetic can regulate brain activity in sub-second scale resolution, it has become widely adopted as versatile tool to integrate with other neuro-technologies including touchscreens.
Numerous examples demonstrate that existing mouse models have not been fully successful in providing accurate information on how pathological changes impact cognition. For example, several mouse models have been generated that express transgenic human amyloid precursor protein (APP)—the precursor of the primary component of amyloid plaques (ß-amyloid peptide)—with familial Alzheimer’s Disease mutations using non-specific promoters. Unfortunately, in most of these models APP is overexpressed in circuits and cells without the spatiotemporal relevance required to model Alzheimer’s disease, leading to overexpression of a number of extra peptides in addition to ß-amyloid in the incorrect cell types and circuits. Similarly, Tau, a protein that when misfolded leads to neurofibrillary tangles, the second main biomarker in Alzheimer’s Disease, has also been overexpressed in myriad ways in mice, without modelling the specific Tau isoforms found in humans. Such models not only inadequately model disease pathology but, because of their severity, increase the likelihood of false positives on cognitive assays.
Thanks to recent developments (such as use of CRISPR/CAS) and multiple efforts from individuals and consortia (e.g Model-AD), next generation mouse models in which specific humanized genes are knocked-in the mouse genome are slowly replacing first generation mouse models. The use of robust, automated, translational and reproducible touchscreen testing of high-order cognitive functions means that these subtler and more realistic mouse models, which may provide more relevant information but are also likely to display subtler cognitive impairments, can now be used. We have access to a number of these next-generation mouse models. See below for a list of all mouse lines available via the Rodent Cognition Research and Innovation Core.
We also emphasise knowledge and data‐sharing practices to create an epistemic community of scientists who share common methodology and are united in the goals of increasing methodological transparency and improving the reliability and reproducibility of research findings. To encourage other researchers to adopt common methodology, we have created extensive resources that facilitate engagement in open knowledge-sharing of our expertise, platform developments and data, aligned with Canada’s Roadmap for Open Science:
www.Touchscreencognition.org serves as a central hub for state-of-the-art know-how about using touchscreen cognitive testing systems. This knowledge-sharing platform provides access to the latest publications, a forum for discussion, training videos and resources such as standard operating procedures (SOPs).
www.Mousebytes.ca is the first open science data repository for mouse cognitive data. This data sharing platform follows the FAIR principles of open science and has received support from and is being linked to the Brain Canada funded Canadian Open Neuroscience Platform (CONP).