Validity & Reliability of Animal Models

What are the three components of an emotional response?

Behavioural, autonomic, and hormonal

What is behavioural response?

Muscular movements and observable motor patterns appropriate to eliciting the situation

What is autonomic response?

Eg sympathovagal activation, which facilitates the behaviour and quickly mobilises energy for vigorous movement

What is hormonal response?

Modulates (reinforces or attenuates) the autonomic responses

Which DSM depression symptoms can NOT be modelled in mice?

• Depressed/irritable mood

• Feelings of worthlessness/guilt

• Thoughts of death/suicide

What can be modelled as endophenotypes in mice for depression?

• Anhedonia - sucrose preference

• Sleep - EEG

• Locomotion and cognition - water maze, Y-maze

Name the behavioural endpoint groups for depression-like states and an example test for each

• Despair - swim/tail suspension

• Apathy - grooming, nest building

• Anhedonia - sucrose preference, cookie test

• Anxiety - open field

• Social aversion - social defeat

How are anxiety and depression related epidemiologically (Kessler 1994)?

High comorbidity: most with an anxiety disorder develop lifetime MDD, and >50% with MDD develop lifetime anxiety disorder (prevalence ~25% and ~17%)

What does the overlap of anxiety and depression mean?

Animal anxiety and depression endpoints are hard to separate

Give the categories of behavioural assays for fear, anxiety, and stress in rodents

• Fear: Pavlovian conditioning, startle

• Anxiety: ethological tests (EPM, open field, light-dark), active avoidance (marble burying), hypnoneophagia, social tests

• Stress: vital signs, plasma hormones

List the four approaches to constructing animal models of neurospsychiatric disorders

1. Genetic

2. Pharmacological

3. Environmental

4. Electrical stimulation/lesions

Genetic approaches to construct animal model

• Selective breeding

• Mutation

• Transgenics

• Viral delivery

Pharmacological approach to construct animal model

Agonists/antagonists

Environmental approaches to construct animal model

Chronic, social, or physical stress

Electrical/lesions approaches to construct animal model

• Brain stimulation including optogenetics

• Anatomical lesions

Main strength and weakness of the GENETIC approach (for constructing animal model)

• Strength: focus on the phenotype/gene of interest

• Weakness: may be only a phenocopy; variable penetrance; human relevance of phenotype hard to establish

Main strength and weakness of the ENVIRONMENTAL and PHARMACOLOGICAL approaches (for constructing animal model)

• Environmental: recapitulates human risk factors, easy, BUT lacks specificity/construct validity

• Pharmacological: temporal/some spatial control BUT common mental disorders don’t involve a single neurotransmitter system

Name experimental triggers for depressive-like states and the validity they relate to

• Acute/chronic stressors

• Early life manipulations

• Olfactory bulbectomy

• Reserpine/tryptophan depletion

• Immune stimulation

• Psychostimulant withdrawal

Triggers relate to construct (aetiologic) validity

What is Unpredictable Chronic Mild Stress (UCMS)?

Weeks of varied, unpredictable mild stressors followed by a test phase for endpoints such as sucrose preference

What are the advantages of simulation models?

• Closer to real disease symptoms

• Potential for selecting novel targets

What are the disadvantages of simulation models?

• More difficult

• Diversity and lack of knowledge of disease etiologies and pathologies

Define face validity

Phenotypic similarity to behavioural and physiological state of the human disorder

Define construct validity

Similarity of underlying causes; same inducing condition, etiology, and neurobiological/genetic mechanisms

Define predictive validity

Similar efficacy of clinically proven treatments

What are four components of predictive validity?

• Therapeutic class specificity

• Dose-dependency and correlation with clinical potency

• Behavioural selectivity/absence of tolerance

• Reliability (intra- and inter-lab reproducibility and replicability)

What is the difference between replicability and reproducibility?

• Replicability = repeating the experiment to confirm the scientific claim with new data (highest standard of evidence)

• Reproducibilty = redoing the analysis with the original data/code (minimum standard)

What is needed for a study to be reproducible?

• Data + metadata

• Code (processing, analytic, presentation)

• Documentation

What can a study report alone do?

Assess study justificatiuon, assess design, understand how experiment was conducted, and assess relevance

What can a study report alone NOT do?

Assess analysis errors, sensitivity to assumptions, or reproduce the analyses

Whre do individual arise, from egg to test day?

• Genotype/sex/mutations

• Epigenetics/imprinting

• Uterine environment

• Maternal programming

• Birth weight, litter, quality of maternal care

• Nutrition, age, social experience

• Test apparatus, lighting, experimenter

What do individual differences after chronic social stress/defeat look like, and why does it matter for data?

• Animals split nto vulnerable/susceptible vs resistant/resilient (eg social interaction ratio <1 vs >1)

• Group means can hide real effects

How does maternal care shape adult stress responsivity (Meaney)?

• High LG: more hippocampal GR expression → better HPA negative feedback → low corticosterone and anxiety

• Low LG: more Nr3c1 methylation → less GR → high corticosterone and anxiety

Effects are long-lasting and can be transmitted

What is the standardisation fallacy?

Standardisation increases reproducibility at the expense of external validity (face/construct), producing artefacts

What did Crabbe et al. (1999) show?

Found lab-specific artefacts (Portlan, Edmonton, Albany) despite rigorous standardisation

What are the advantages of standardisation?

• Less variability

• Sensitivity

• Precision

• 3R (reduce)

• Internal validity

What are the disadvantages of standardisation?

• Results valid only for narrow conditions

• Lab-specific artefacts

• Lower external validity

• May miss effects

What are the advantages of biological variability?

• Reflects real populations

• Improves external validity and translational relevance

• Lets you study individual differences (vulnerable vs resilient)

What are the disadvantages of biological variability?

• More noise

• Needs larger samples to detect effects

What are the proposed solutions of standardisation (heterogenisation)?

• Genetic: outbred/feral, multiple inbred stains, KOs on multiple backgrounds

• Environmental: independent replicates, multi-centre studies, systematic variation of factors