Research Methods PDF

Research Methods Research Methods How do we study behavioral neuroscience? Cognitive & behavioral models/testing Computational models Molecular biology: in vitro and in vivo Scanning and neuroimaging of the living brain Postmortem studies...

Research Methods Research Methods How do we study behavioral neuroscience? Cognitive & behavioral models/testing Computational models Molecular biology: in vitro and in vivo Scanning and neuroimaging of the living brain Postmortem studies Family studies: genetics Medical studies Animal models Research Methods Spatial resolution: how precisely can you measure something in space; ability to distinguish two points from each other increasing spatial resolution Temporal resolution: how precisely can you measures something in time; ability to distinguish events happening in time Levels of Analysis Reductionism: idea that mind is reducible to brain activity; brain reduced to circuits, then reduced to cellular, molecular, atomic physics,… Experimental Design 1) Define your question(s) How do we localize a cognitive process to a brain region? 2) Which species/model is best to use? Animal, human? 3) Which method is appropriate to answer your question? Know how the method works Understand its advantages and limitations Experimental Design Variable: some measurement that we are making; could be a brain recording, hormone measurement, questionnaire people fill out, anything! Fun going in a pool Correlation: looking for a relationship Chess skills between two things Height E.g., height and age are correlated with each other When people get older, they also get taller. Age Age Height Each point is a person Experimental Design Confounding Variables Ice cream sales and drowning are correlated with each other! Does this mean that selling ice cream kills people?! There must be a third variable that makes sense of all this. This third variable that is a better explanation for the relationship is called a confounding variable. Experimental Design Confounding Variables Ice cream sales and drowning are correlated with each other! Nice Does this mean that selling ice cream weather kills people?! There must be a third variable that makes sense of all this. This third variable that is a better explanation for the relationship is called a confounding variable. Buy Go to the pool ice cream X and drown Correlation analysis Experimental Design How do we localize a cognitive process to a brain region? Contrast: Isolating a variable minus What does this contrast show? Experimental Design Contrast: subtracting one condition from another to isolate a single variable Need to think about all the different confounding variables and try to remove them Find a brain region that goes up little by little as the person shows more fear In the previous contrast, we wanted to find just “fear”, but instead we also found faces, high contrast, visual content, attention, and more! Experimental Design Contrast: subtracting one condition from another to isolate a single variable Need to think about all the different confounding variables and try to remove them Find a brain region that goes up Amygdala little by little as the person shows activates to more fear emotional faces In the previous contrast, we wanted to find just “fear”, but instead we also found faces, high contrast, visual content, attention, and more! Wang et al., Nature Comm (2017) Experimental Design Experiment: administering an intervention that changes something Change something -> see what happens Independent variable: the variable that the experimenter changes to see what happens Dependent variable: measure what happens when you changed something This variable is “dependent” on what you changed Experimental Design We want to limit the amount of confounding variables so we need a controlled setting There should only be ONE variable that is changing; this is the independent variable. Control group: another group that receives a placebo or a fake intervention just to make totally sure that the experiment itself doesn’t change the dependent variable Experimental Design You always need to be careful about confounding variables What if the participants figure out that they are in the control group? This means that they could lose motivation to continue the experiment What if the experimenter finds out that the person is getting the fake intervention and starts acting different around them? Experimental Design Blinding: intentionally design experiment so participants do not know that they are in the real group or the control group Double-blinding: make sure that experimenters do not know which participants are in the real group or the control group Blinded analysis: analyst does not even know during the analysis of the data which group is which until the very very end Experimental Design The placebo effect is strong; very often, placebos outperform no treatment at all Placebo surgeries have sometimes outperformed the surgery! Effect is dependent on how convinced one is that they are getting good treatment Drugs needs to outperform a placebo to be considered effective treatments Lesser known Nocebo Effect: you make yourself worse! Experimental Design Between-participants: one group gets the real intervention and another group gets the control (fake) intervention Also called “parallel arms” because there are two arms outstretched that do not overlap with each other Experimental Design Within-participant: each person gets the real intervention and the fake intervention Also called “crossover study” because there are two paths that cross over each other. Either you get Arm-A, then Arm-B OR Arm-B, then Arm-A Experimental Design Somatic intervention: change their body Behavioral outcome: see if behavior changed Behavioral intervention: change their behavior Somatic outcome: see if their body changed Experimental Design From Correlation to Causation An example First, we identify a correlation between two variables. 1. Serotonin level in the blood is negatively Second, then we design an experiment to correlated with depression severity. test whether there is a causal relationship 2. Design a parallel arm between-participant between these variables by delivering an experiment to give half of the people with intervention to change one of them. depression serotonin in a pill and the other Third, check that your intervention engaged half get a placebo. the target. Did you change the thing that you 3. Check that the pill increased serotonin in targeted? the blood. Fourth, see if the dependent variable 4. Investigate whether people showed changed as a result. reduced depression severity. This process builds causal evidence. Research Methods Invasive: technique requires you to enters the body Greater precision Limited to animal models or medical need Noninvasive: technique does not enter the body Ideal for routine human research Less side effects Not always a distinction between animal and human research In vivo (“within life”): performing an experiment in an intact, living organism In vitro (“within glass”): experiment performed on cultured cells or isolated molecules of DNA, RNA, protein ex vivo: a section of a living organism is taken (e.g., tissue biopsy) Research Methods Importance of human and animal research: 1. Underlying mechanisms of behavior are similar across species 2. Interested in animals for their own sake 3. What we learn about animals informs us about human evolution Universities and research institutions in the US are required to have an Institutional Animal Care and Use Committee (IACUC) and an Institutional Review Board and Protection of Human Subjects that oversees and determine acceptable procedures Experimental Design The human techniques are lower spatial and temporal resolution… Why would you study humans? 1. Higher-order cognition: language, decision- making, emotion / social processes 2. Different brains might be different across species 3. You can measure the whole brain with MRI / EEG; you might miss something if focused on a particular region or handful of regions Research Methods Imaging the brain Structural imaging: looking at the fixed anatomy of the brain/cells Functional imaging: looking at the dynamics of the brain in action X-Ray Based Imaging Techniques (structural) Contrast x-rays Inject radio-opaque material into structure of interest E.g., Cerebral angiography: dye in carotid artery Computed tomography (CT) 3-dimensional, but not high-resolution Identify brain bleeding, tumors, traumatic injury Blood Tumor Bullet fragment Magnetic Resonance Imaging (MRI) 1946: Nobel Prize Medicine Applied to Medicine in 1970s, first clinical scan early 1980s Utilizes hydrogen atom ~70-80% of all atoms in the body are hydrogen H2O the most common (~60% body weight) Magnet Strengths are measured in Teslas: 1.5, 3.0. 4.0, 7.0 3T = 3 Tesla 1 Tesla ~ 20,000 times earth magnetic field Magnetic Resonance Imaging (MRI) How MRI Works 1. Use a high voltage current in a superconducting copper coil at 2 degrees above absolute zero to create an extremely strong magnetic field. 2. All of the protons (hydrogen) in the body align to the magnetic field when you enter the room. Magnetic Resonance Imaging (MRI) How MRI Works 3. Shoot a radio frequency pulse at the brain to rotate the protons in the magnetic field. 4. The protons rotate back to the main field. However, the magnetic properties of the protons changes based on the tissue, so it takes a different amount of time to go back to the magnetic field. Magnetic Resonance Imaging (MRI) How MRI Works 5. Take a picture of the brain one slice at a time. Each slice takes ~two seconds to collect! So it takes 5-10 minutes to get one 3d image. 6. However, the images are collected in “k-space” which is the Fourier domain of space. Apply an inverse Fourier transform to recover the image. Magnetic Resonance Imaging (MRI) (structural) Structural/anatomical MRI): Images are measurements of the waves emitted by hydrogen atoms after they are placed in a magnetic field spatial resolution is higher than CT; due to neural structures differ considerably in their density of hydrogen atoms Diffusion tensor imaging (DTI): Images axonal tracts (rather than cell bodies); identifies connections between different regions. Based on rapid diffusion of water molecules long white matter tracts (i.e., myelin) Functional Magnetic Resonance Imaging (fMRI) Maps blood flow and oxygenation (blood-oxygen level dependent; BOLD signals) to identify parts of brain active during particular tasks After a brain region activates, it requires blood to re-oxygenate High spatial resolution (~2mm) Too slow to catch neural activity (temporal resolution ~6-10 sec) Ideally, contrast trial types to localize area(s) of interest Positron Emission Topography (PET) Functional imaging, invasive?? Radioactivity-Based Techniques Tag a molecule with a radioactive tracer Intravenous injection while in scanner; radioactive tracer enters brain and gets processed by certain neurons Tracer undergoes radioactive decay and emits photons; photons interacts with electrons to release gamma rays at 180 degree angle PET scanner measures these gamma rays at a detector and localize where in the brain they originated Positron Emission Topography (PET) Alzheimer’s Dementia Non-Alzheimer’s Dementia Tag specific molecules E.g., beta amyloid plaques are linked to Alzheimer’s disease E.g., changes in dopamine production in striatum with substance use disorder Amyloid Beta Plaque Signal Imaging Cellular Structure Histology: the study of the structure of tissue Staining: use chemicals that have different affinities for different parts of cells Nissl Stain: stain that outlines all cell bodies; dyes attach to the RNA which is centered on the nucleus (center of cell body) Golgi Stain: only a small set of cells are stained but the full cell is outlined Nissl Stain Golgi Stain Tracing connections between neurons Tracer: chemical that emits light so that you can take a picture and see it in the picture Tract tracers: while the animal is still alive, give a part of the brain some “tracer” and then it will be transported around between neurons; then, sacrifice and image those connections anterograde trace: stains cells from the soma to the axon terminal retrograde trace: taken up by the axon terminals and stains the soma Electrophysiology Electrophysiology: recording (and sometimes manipulating) electrical activity in neurons and brains Patch clamp: suction cup to one neuron and record electricity. Microelectrode (array): one (or many) Patch clamp electrical recording sensory; population Microelectrode array of neurons. Electrocorticography (ECoG): sensors placed directly on or in the brain; specific area of the brain Electroencephalography (EEG): sensors placed on the scalp; limited spatial resolution Electrocorticography Electroencephalography Electrophysiology Invasive electrophysiological recording methods Electrophysiology Scalp electroencephalography (EEG): measured by scalp electrodes Event-related potentials (ERPs) Sensory evoked potentials Signal averaging: increases signal- to-noise ratio P300 wave Electrophysiology Magnetoencephalography (MEG): measures magnetic fields in brain instead of electricity Magnetic fields pass right through the scalp, perpendicular to electric fields Cons: expensive; you need superconducting quantum interference devices (SQUIDs) that are 2 degrees of absolute zero in order to measure subtle magnetic fields Pros: higher spatial resolution than EEG Lesions Lesions are areas of brain that are damaged by disease or injury Traumatic brain injuries, stroke, or tumors can result in lesions that affect behaviors In 1860s, Broca studied patients with difficulties producing speech Cluster symptoms and cognitive changes based on damage location; localization of function Caveats to Lesion Studies Lesions are rarely small enough and specific enough to affect just one brain area. Lesion may spare enough of the tissue that the function is preserved, making it seem that that brain region is not involved in the behavior Researchers must be careful to characterize deficits precisely Lesions Ablation studies: destroy the part of the brain and that was found to be important for a particular behavior Muscimol inactivation: inject this compound in a region of the brain and shut it down temporarily Stereotaxic Surgery: employs stereotaxic atlas and instrument Allows accurate placement of lesions, probes, electrodes, or other instruments Reference point used is bregma Gene Techniques Gene knockout: remove a particular gene from the genetic code of a mouse; mouse then grows up without this gene Wild-type: a control group of mice that do not have any genetic manipulation After finding a correlation between a gene and a behavior, you can “knockout” that gene and see if the behavior changes Gene knock-in: Transgenic mice Genetic information from a different species is A laboratory mouse in which a gene affecting implanted hair has been knocked out (left), is shown next to a normal lab mouse (right). Gene editing: CRISPR/Cas9 method shows exceptional promise Cas9 protein is linked to guide-RNA Integrated into a virus Injected into an organism Genetic Engineering The Central Dogma How do we get a neuron to artificial generate a certain protein? Viral Infection: Viruses highjack the protein creating system of the cell Rabies and herpes viruses are modified to express certain proteins of interest; can be designed to target specific neurons Gene Editing: CRISPR-Cas9 is a technique to insert DNA into DNA sequence Add or subtract some part of an animal’s DNA and then let them grow up Genetic Engineering Brainbow: Insert genes that make neurons express many different fluorescent proteins Fluorescent means it emits color Remove all lipids from brain and then take a picture Illustrates many neural connections simultaneously Tract Tracers + Gene-Editing Can target a specific type of neuron with a “viral vector” - then, change that neuron’s protein expression to create tracer Now that tracer spreads to its connected neurons; can also make this go “forward (antegrade trace)” across post-synapse or “backward (retrograde trace)” across pre-synapse Can figure out which types of neurons are connected to which others Can use two different colors with two different traces and see where the two colors are co-expressed. “Red + green = yellow” Imaging Proteins in Neurons Locating Neurotransmitters and Receptors in the Brain Immunocytochemistry Create antibodies to desired neurotransmitter or receptors Label antibody with dye or radioactivity This section through a rat’s pons reveals noradrenergic Look for specific neuroproteins in brain slices neurons that have attracted the antibody for dopamine- beta-hydroxylase, the enzyme that converts dopamine to norepinephrine. In situ hybridization Locates peptides and proteins in the brain Labeled hybrid RNA that complements mRNA is administered They bind and allow scientists to locate the neuroprotein of interest Use with microscopes (e.g. fluorescence microscope) Brain Stimulation Researcher stimulate or increase the activity in a brain region and observe how behaviors change Penfield and Jasper (1950’s) used electrical stimulation of brain regions to map the sensory and motor areas Patients undergoing brain surgery for epilepsy treatment Not under anesthetics; can respond Brain Stimulation Electrical Stimulation: just as electricity can be recorded at many different levels, you can stimulate electrically at each of these levels Transcranial electrical stimulation: non-invasives stimulation through pads placed on the scalp Deep brain stimulation: implanted electrodes for medical purposes in humans and research purposes in animals Electrical microstimulation: stimulation at the microscopic level of populations of neurons Brain Stimulation Transcranial magnetic stimulation (TMS): uses electromagnetic coils to activate neurons in a particular region of the brain Brief magnetic pulse that is converted into electricity in the brain and directly activates neurons Make your hand twitch from stimulating your brain – used for calibrating the stimulator Repetitive transcranial magnetic stimulation (rTMS) can be used to treat depression Brain Stimulation Transcranial Focused Ultrasound Stimulation: delivers mechanical vibrations to the brain and activates the targeted area Sound waves measure changes in blood volume The same ultrasound as in pregnancy! Very specific focal targeting New tool with limited usage at this time Optogenetics Insert light-sensitive ion channels into neurons (genetic editing; opsin genes) Then, shine a laser on those ion channels to open the channels One ion channel activates the cell and another inhibits the cell; hyperpolarize or depolarize cells We now have an ON/OFF switch for neurons! Optogenetics Video Behavioral Paradigms Paradigms for the Assessment of Species-Common Behaviors Open field test Tests of aggressive and defensive behaviors Tests of sexual behavior Traditional Conditioning Paradigms (Pavlovian conditioning, operant conditioning Radial arm maze, Morris water maze Studies rat spatial ability Neuropsych Assessments Intelligence (Wechsler Adult Intelligence Scale) Memory Language tests Critical-Thinking Scientific method provides a systematic way to 1. Observation: Careful, repeated Findings are study a process and avoid biases and heuristics. reviewed by peers measurements using standardized Observations Make observations about world and replicated or refuted by further techniques lead to research questions. Develop a hypothesis to explain observations experiments. Generate testable predictions about 4. Experimental testing of hypothesis predictions: Overcoming bias: Standardization of 2. Development of a hypothesis: a requires careful Perform experiments to test predictions measurement and measures; Disconfirmation of proposed explanation for the comprehensive Results of one experiment help to refine hypothesis; Peer observations in efforts to control for review of findings; question; a good any additional hypotheses for the next experiment factors that could Replication of results; Progress hypothesis should be falsifiable. influence the Experiments are repeated and extended to show by iterations observations. they are reproducible. Results are subjected to independent peer review Experiments are 2. Generation of specific and testable predictions: if Competing hypotheses lead to reduce the biases that may have been designed to test the predictions of the our hypothesis X is true, to competing predictions. introduced by the experimenter hypothesis against alternatives. then we ought to be able to observe other phenomena A, B, and C. The Payoffs of Cognitive Neuroscience Healing the Disordered Brain E.g., techniques such as rTMS and deep brain stimulation can help alleviate symptoms of conditions such as depression, Parkinson’s disease, and obsessive-compulsive disorder Enhancing Human Abilities Understanding how humans make decisions can provide insight into how to make better decisions Brain interface devices (cochlear implants, implanted electrodes to enable paralyzed patients to move devices outside their own body) can restore lost functions to individuals Blueprints for Artificial Cognition Improve the abilities of technology; e.g. speech recognition, locomotion, and object recognition Artificial “neural networks” are used to recognize patterns of brain activity in brain imaging studies Much remains to be done to find artificial equivalents to much human cognition Brain-Compatible Social Policies Eyewitnesses are very compelling in courtroom, but are notoriously unreliable; cognitive neuroscience can help understand and explain the factors that influence the memory of such witnesses. Medical practices (e.g., Brenda Milner and patient HM informed neurosurgeons) ~25% of American prison population has a mental illness Reminders Final Exam is actually Tuesday April 29, 7:30-9:30am!!