Thomas Carlson lectures Flashcards
Describe the history of brain measurement methods
Year: Pre 1900 Person: William James Description: -Accidental experiments ---Placement of bullet shots in different parts of the brain determined future abilities if patient survived -Medicine and philosophy -Birth of psychology
Year: 1924
Person: Hans Berger
Description: Collects first EEG recording
Year: 1928
Person: Edgar Adrian
Description: First recording of single unit activity
Year: 1968
Person: David Cohen
Description: First MEG recording (direct measure)
Year: 1973
Person: Hoffman, Ter-Pogossian and Phelps
Description: First PET scan
Year: Late 1970s
Person: Lauterbur and Mansfield; and Damadian
Description: Develop MRI technology
Year: 1980
Person: Merton and Morton
Description: Stimulate motor cortex with TMS
Year: 1991
Person: Belliveau et al. + Ogowa et al.
Description: Develop fMRI
What are newly developed techniques for measuring brain activity/analysis of this data and how did it come about?
-Brain science is multidisciplinary • Cutting edge techniques o Optogenetics Shine light onto neurons that have been genetically modified to increase their activity of decrease their activity o Two photon imaging Can only be done in animals Real time imaging of cellular activity • New and upscaled analysis methods have allowed for further learning about the brain o Machine learning Reinforcement learning Supervised learning • Classification • Regression Unsupervised learning • Clustering • Dimensionality reduction
Is there a perfect methods to measure the brain? Why/why not?
• No method is perfect o Have to consider spatial and temporal dimensions, as well as invasiveness o The right method depends on the scale of the question Components of a neuron • E.g. Physiology Neurons • E.g. Physiology Neural networks • E.g. Physiology and arrays Brain areas • E.g. fMRI Brain networks • E.g. fMRI
What are the Barlow’s 5 dogmas and what are they arguments for?
• An argument for direct measures and observation of single neurons through single-unit recording- Barlow (perception, 1972)
o To understand nervous function, one needs to look at interactions at a cellular level, rather than either a macroscopic or microscopic level, because behaviour depends upon the organised pattern of these intercellular interactions. This dogma is possible as single neurons have diverse and highly specific responsiveness to sensory stimuli, and are astonishingly reliable
o The sensory system is organised to achieve as complete a representation of the sensory stimulus as possible with the minimum number of active neurons. This dogma is possible as, at higher levels, fewer and fewer cells are active, but each represents a more and more specific happening in the sensory environment.
o Trigger features of sensory neurons are matched to redundant patterns of stimulation by experience as well as by developmental processes (including genetics)
o Perception corresponds to the activity of a small selection from the very numerous high-level neurons, each of which corresponds to a pattern of external events or the order of complexity of the events symbolised by a word. However, not every cortical neuron’s activity has a simple perceptual correlate
o High impulse frequency in a given neuron corresponds to high certainty that the trigger feature is present.
What is an argument for measuring neural networks instead of single neurons?
• An argument for neural networks
o Instead of looking at single neurons, scientists should be looking at how neurons interact with each other and their activity
o Avoids the grandmother cell effect
Don’t need one cell to encode a person- encode a person through a joint activation pattern across multiple neurons
What are the two types of brain measures? Define and give examples of technology used for these brain measures
• Direct measures
o Direct measures relate directly to neuronal activity
E.g. single unit recordings, EMG, EEG
• Indirect measures
o Indirect measures use a conduit to access neural activity
o They are associated with neural activity (although not necessarily linearly) but measure other factors
E.g. fMRI measures changes in blood oxygenation
What techniques are used to measure brain activity?
o Single unit recordings o Electroencephalography (EEG) o Magnetoencephalography (MEG) • Local field potentials • Optogenetics
What are single unit recordings?
Measures action potentials for individual neurons
What are the two types of single unit recordings, and their advantages/disadvantages?
Types:
• Extracellular-
o Doesn’t penetrate cell body of the neuron: instead, records extracellular fluid properties
o Recordings can occur for months before the cell starts dying
o Reads basic components of an action potential
o High spatial and temporal resolution
• Intracellular-
o Penetration in neural cell body
However, this means that the cell begins dying as soon as penetration occur, so recording can only occur during a limited time frame
o Records from exactly one neuron
o Can record subthreshold potentials
What is the procedure for single unit recordings?
- Record activity for a stimulus
- Identify spikes (spike sorting)
- Measure spiking activity over time
- Repeat for multiple trials
What are properties of a good extracellular electrode? Give an example of one
o Properties of a good extracellular electrode
Ruggedness to pass through neural tissues
Stability for accurate localisation
High signal-to-noise ratio (low electrode noise)
High selectivity
o Examples are platinum black-plated platinum or stainless steel electrodes (which do not require a trade-off between high signal-to-noise ratio and selectivity as other electrodes do)
Describe how spikes are produced in single unit recordings, the theoretical recording of an isolated axon (and why it is so) as well as why theoretical values do not always match reality
o The action potential (or spike) recorded with an extracellular microelectrode is produced by currents that are induced to flow in the extracellular space around an active neuron
Theoretically records a triphasic waveform from the isolated axon
• As the action potential approaches the region underneath the electrode, the electrode sees a positive potential relative to a distant indifferent electrode
• When the action potential reaches the membrane underlying the electrode, the electrode records a negative potential
• As the action potential continues down the axon, the membrane under the electrode once again records a positive potential
However, as the extracellular space does not have uniform, low resistance, the measured potential is not always accurate
Why is there a need to separate and identify spikes in extracellular single cell recordings? How is this done?
o There is a need to separate and identify spikes as extracellular recordings may pick up more than one neuron
Multidimensional spike sorting allows confident identification of an individual cell, as well as simultaneous recording of ensembles of cells, with on-line and/ or off-line separation based on waveform parameters
What is needed if the cell that is recorded during single unit recording is inactive?
o Sometimes need search stimulus activation (electrical of physiological stimulation) of neuron of interest for the extracellular electrode to pick up the spikes
Are single unit recordings often done in humans?
Not easily done in humans- mostly done during epileptic surgery
What is electroencephalograhy (EGG) and how does it work?
EEG measures the summation of electrical activity on the scalp, primarily derived from post-synaptic activity around the dendrites of pyramidal neurons in the cerebral cortex
• Pyramidal neurons are found in the most superficial layer of the brain and are spatially aligned; thus, their activity is synchronous
o This produces a larger signal that can be measured by the EEG
• Summation of the dipoles created by many neurons is what is detected by the EEG
o When the pyramidal neurons fire, they create currents (due to depolarisation and repolarisation at different areas of the pyramidal neuron- follows the action potential)
Is EEG safe for humans?
Safe for humans
• However, electrocorticography and intracranial electroencephalography are invasive
What are the similarities between MEGs and EEGs?
MEGs and EEGs are both direct measures of neural activity and both have excellent temporal resolution (milliseconds).
What are the differences between MEGs and EEGs?
However:
• They measure different signals
• MEG has an advantage in that magnetic fields are not affected by the skull. However, EEGs are affected by skull and tissue interference
o Hence, MEG has slightly better localisation
• Different sensitivity- EEG can measure deep sources in the brain, but MEG is not as sensitive to deep sources (only measures activity at the surface of the brain)
o Electrical signals do not drop out as fast throughout distance compared to magnetic signals
• MEG ($1000) costs a lot more than EEG ($15)
o Sensors in MEG need to be bathed in liquid helium, which is very expensive, and the MEG apparatus needs to be in a room that blocks out interfering magnetic signals
What is a common disadvantage of MEGs and EEGs?
Both have a disadvantage:
• Source localisation can be difficult as, although recordings are done at the scalp (which is a 2D surface), the signals are coming from the brain (which is a 3D object)
• However, there is a tight link between the activity measured by MEG, EEG and single unit recordings, suggesting that spatial and temporal resolution of both techniques is not abysmal
What do local field potentials measure?
Measures electric potential in the extracellular space around neurons- reflects changes in synaptic activity
What are the stages of an action potential?
• Stages of an action potential-
o Resting potential
The voltage across the membrane is about -65 mV at resting potential
o Gets triggered
Near the terminal bouton, there are voltage-gated calcium channels
When an action potential fires and gets down to terminal, action potential opens calcium channels which will call the synaptic vesicles to dock and release their neurotransmitters
Neurotransmitter binds to ligand-binding neurotransmitter receptor which opens up sodium channels which starts making the cell depolarised
When the threshold is surpassed (-55mV), an action potential occurs
• An action potential is ONLY caused when the depolarization of the membrane is beyond the threshold
• Threshold-the membrane potential at which enough voltage-gated sodium channels open so that the relative ionic permeability of the membrane favors sodium over potassium.
o Rising phase
Sodium channels open and inside of cell becomes more positive as sodium enters in the cell
o Overshoot
Voltage is about +35 mV- the part where the inside of the neuron is positively charged with respect to the outside
o Falling phase
Potassium channels open and inside of cell becomes less positive, returns to negative
Sodium channels close
o Undershoot
Rapid depolarisation causes the inside of the membrane to be more negative than the resting potential
o Absolute refractory period
Sodium channels inactive when the membrane becomes strongly depolarised. They cannot be activated again, and another action potential cannot be generated, until the membrane potential becomes sufficiently negative enough to deinactivate the channels (usually 1msec)
o Relative refractory period
The membrane potential stays hyperpolarised until the voltage-gated potassium channels close. Therefore, more depolarizing current is required to bring the membrane potential to threshold
o Restoration of the resting potential
List indirect measures of measuring/determining brain activity?
• Brain damaged patients o Often from bullet holes and accidents o Stroke, concussions and brain injury • Behavioural approaches o Qualitative methods o Questionnaires o Psychophysics • X-rays • CT scan • MRI • Diffusion Tensor Imaging • Metabolic methods o PET (Positron Emission Tomography) imaging o fMRI (functional MRI) o Radioactive tracers (only in animals) • Optical imaging (Only in animals) • Transcranial magnetic stimulation (TMS) • Clarity
What people are examples of how brain damage and injury can be used to learn about brain functions? Describe what happened to these people and what were learnt from them
Phineas Gage
• Railroad spike went through frontal lobe- learned that frontal lobe is important for higher cognition/decision making
• Showed lack of ambition and became aggressive
H.M
• Treated with brain surgery for epilepsy- removed both his hippocampi
• Learned that the hippocampus is critical for forming long term memories
Victor Leborgne (Tan)
• Had a stroke that affected Brocas area (critical for the production of speech)
• Still understood words and instructions
NFL players who headbutt each other at full speed often experience depression, high suicide rates
• Tied to repeated concussions that players had (CTE syndrome)
What was an initial disadvantage of studying brain functions using brain injury/damage and how was it done?
o Would often have to wait for subjects to die of natural causes before observing their brain
Behaviour+ Anatomy+ Patience= knowledge
What is psychophysics? Give examples of such techniques
Precise manipulations of stimuli to understand how the brain encodes information
• Adaptation
• Psycho-anatomy
• Divided visual field presentation
Describe what the disadvantage of X-rays in in elucidating brain function
o Hard to fully observe inside the skull
Describe why CT scans were so revolutionary in elucidating brain function
o Can look at the brain in 3D- easy to look at inside structures of the brain
o No longer have to wait for patient death to observe their brain anatomy
How does MRI work?
o Each atom contains protons which are positively charged and spinning around an axis
o This spinning charge creates a small magnetic field
o An MRI machine is effectively a large magnet so that when an individual is placed inside the magnet, their magnetic protons line up along the longitudinal axis with the MRI magnetic field
o A short strong electromagnetic pulse (radiofrequency pulse) is applied to disturb the spinning protons- this results in less of them spinning in the same axis as the MRI magnetic field and more spinning in a transverse plane
o Once the radiofrequency pulse is turned off, the protons return to their original direction along the MRI magnetic field- longitudinal magnetization gradually increases until it reaches the same point where it was before the radiofrequency pulse
The rate of this return is called the longitudinal relaxation time or T1 time
The rate of decrease of the transverse relaxation is called the transverse relaxation time or T2 time
The T1 and T2 relaxation times are different and are independent processes
• Rate of decay gives information on different kinds of tissues
How does MRI recognise different tissues?
Different tissues have different T1 and T2 relaxation times
Water has a high T1 and T2 relaxation time whereas fat has lower T1 and T2 relaxation times
By applying a series of radiofrequency pulses the MRI can distinguish between different tissues that have different relaxation rates
How are MRI scans prepared for analysis?
Taking multiple images and averaging them produces greater contrast between gray matter and white matter
Even higher spatial resolution and grey/white differentiation is possible with longer scan times and higher field strengths
What is tractography?
tracing tracts by visualising fibres
What is the basis of DTI?
o DTI- based on use of diffusion of water molecules to generate different pixel intensities that can contain information about direction of water movement
What is mean diffusivity?
o Mean diffusivity-
A measure of the average molecular motion independent of any tissue directionality.
It is affected by cellular size and integrity
Anything that changes the physical structure of the area will alter mean diffusivity
What is the concept behind diffusion tensor imaging?
Diffusion of protons depends on freedom of movement in tissue
• Protons move differently in different brain compartments
• Membranes restrict movement
In fibre tracts, combination of movement in axonal membrane and surrounding oligodendrocyte membrane (myelin) restricts movement of protons
Restricted proton movement can be discerned in MRI
o The diffusion matrix (tensors) can be used to generate a 3D image of the tracts
Outline squares in the matrix where protons can’t move much: these will be the fibre tracts
In the myelin of oligodendrocytes, there is directionality: can tell moving from anterior to posterior/ superior or inferior
• Hence, DITs can show which direction the fibres are projecting in and the fibre tracts in an area
Describe the properties of protons in CSF that are important in diffusion tensor imaging
o Isotropic
Moves in all directions
o Water
o High diffusivity
Describe the properties of protons in grey matter that are important in diffusion tensor imaging
o Isotropic
o Low diffusivity (move in any direction but constrained-in smaller space)
o Lots of water in cells but also lipid, cell membranes
Describe the properties of protons in white matter that are important in diffusion tensor imaging
o Anisotropic
Constrained direction- can only move in very narrow space along longitudinal axis of axons
o High diffusivity
o Mostly myelin but also axonal membranes
What is fractional anisotropy?
o FA fractional anisotropy- scalar value between zero and one that describes the degree of anisotropy of a diffusion process- a value of zero means that diffusion is unrestricted and value of one means that diffusion is fully restricted
What is the basis of positron emission tomography?
PET measures positrons following injection of a radioactive tracer containing a positron-emitting isotope into the blood stream
More glucose uptake= more gamma radiation= greater image intensity
What is the process of PET?
A commonly used PET tracer is a labelled form of glucose- fluoreoxyglucose (FDG) which is a glucose analogue
• Although can also tag water for blood flow or receptors for neurotransmitter release
FDG is taken up by high-glucose-using cells such as neurons, where it is then phosphorylated which prevents the glucose from being released again from the cell
The phosphorylated FDG cannot move out of the cell before its radioactive decay which has a half life of 110 minutes
When radioactive nuclei decay, they emit positrons
These positrons collide with a nearby electron resulting in the emission of 2 gamma rays in opposite directions
Gamma rays are sensed by detectors surrounding the brain
Original gamma ray emission location is then reconstructed
• 4-5 minutes
What is the principle behind fMRI? Explain
Most common form of fMRI and relies on the finding that neuronal activity is associated with changes in regional blood flow
T1 relaxation- not related to changes in haemoglobin
T2 relaxation- affected to changes in haemoglobin
• Magnetic susceptibility of deoxygenated haemoglobin is about 20% greater than haemoglobin
• Magnetic susceptibility of deoxygenated haemoglobin affects rate of T2 relaxation
o Magnetic properties of a blood cell (haemoglobin) depends on whether it has an oxygen molecule
With oxygen-> zero magnetic moment
Without oxygen-> sizeable magnetic moment (paramagnetic)
What is the process of fMRI and why does this process work?
• fMRI is a relative measure so you need to compare signal changes relative to a baseline period during the same scan
o Can only be used to measure evoked activity
• When blood vessel is full of deoxyhaemoglobin, disrupts magnetic field
• When blood vessel is full of oxygenated haemoglobin, there is no disruption of magnetic field
• In a resting neuron, the amount of oxygenated haemoglobin entering the capillary bed is about equal as the amount of deoxygenated haemoglobin leaving the capillary bed
• In an activated neuron, there is oversupply of oxygenated haemoglobin and hence leaving capillary body has more oxygenated haemoglobin than it would have at rest
o Oversupply results in less overall distortion of magnetic field-> measured as an increase in blood oxygen level dependent signal
• However, increased blood flow to meet oxygen demands in activated neurons slightly delayed (4-6 seconds)
Explain the shape of the fMRI BOLD response graph
- Initial dip- reduction in the amount of oxygen in that brain area
- Positive BOLD response- brain detects that the area of interest is using many resources and rushes new fresh blood to brain area
- Negative BOLD response- stimulus is shut down-> blood is not urgently provided to area of interest anymore
What are the advantages of fMRI?
- Good spatial resolution
- Non-invasive, not requiring injection of radioactive materials like PET. Subject can be repeatedly scanned
What are the disadvantages of fMRI?
- Noisy
- Susceptible to motion artefacts
- Areas near bone tissue interfaces are susceptible to artefacts
- Metal implants can be dangerous
- More susceptible to neural inputs than outputs
- Bad temporal resolution
Describe how imaging using radioactive tracers is done in animals
o Radioactive tracers (only in animals)
Radioactive tracer given
Sacrifice the animal and flatten the desired section
Look at path that the radioactive tracer has gone
Describe optical imaging as a way to measure brain activity. Can it be done in humans?
• Optical imaging (Only in animals)
o Put a camera directly on surface of the brain and film with a camera
o Look at ratio of blue: red light
If a stimulus is shown, the area of the brain activated due to this stimulus will reflect blue light more than red light
Describe how transcranial magnetic stimulation (TMS) works and what it allows
• Transcranial magnetic stimulation (TMS)
o Uses a magnet in a shape of a coil
Contains a couple of copper windings which passes a strong electrical current, producing a strong magnetic field coming out of the coil
o Place coil on top of people’s heads employs magnetic pulse through the skull of the patient, interfering with brain activity directly underneath the coil
o By targeting different areas in the brain, can look at different brain activities
o Finds if a brain area is causally involved with a function by magnetically interfering with/stimulating that brain area
Describe Clarity as a way to image the brain and how it is performed
o Make the brain transparent and use labelling molecules to label areas of interest
o Can be done on the whole brain
o To make brain transparent, first add a hydrogel mesh (to hold all components together), then remove the fat
Who was Claude Shannon and what important contribution did he make?
• Claude Shannon-
o Key character in developing and quantifying the concept of information
o Worked on missile control systems
o Worked for a phone company
o Proposed the information theory
o Found that compression of information makes for more effective communication
o Applied concept of entropy-the amount of disorder- to information processing
o Provided framework for thinking about the brain
What is the information theory and why was it originally proposed?
o The information theory- studies the quantification, storage and communication of information
Originally proposed by Claude Shannon to find fundamental limits on signal processing and communication operations such as data compression
Describe the relationship between entropy and compression. Give an example
The higher the entropy, the more difficult the compression
• E.g. an image of an empty blue sky would compress to an extremely small file whilst an image of static would result in a very big file size
o JPEG is such a compression tool
How is the information theory applicable to the brain? Which of Barlow’s dogmas does this match with?
• Applicable to the brain- the brain wants the smallest number of neurons firing such that it can code important information in an environment without having to use many neurons (which would use a lot of metabolic energy)
o Matches with Barlow’s second dogma- The sensory system is organized to achieve as complete a representation of the sensory stimulus as possible with the minimum number of active neurons
What is the purpose of the Bayesian framework?
• Bayesian framework
o Allows the use of previously obtained knowledge/beliefs to help the calculate the probability of a related event
What percentage of our body weight does the brain consist of?
2%
What percentage of our energy does the brain consume?
20%
Define the Bayes’ theorem and what each letter describes
o Bayes’ theorem is defined as P(HIE)= ((P(EIH)*P(H))/P(E)
Where H is the hypothesis
Where E is the evidence
Where P(HIE) is the conditional probability that hypothesis occurs given the evidence
Where P(EIH) is the likelihood of evidence
Where P(H) is the marginal probability that hypothesis is true
What is Bayesian inference?
o Bayesian inference is the process of deducing properties about a population or probability distribution from data using Bayes’ theorem
According to the Bayesian framework, what two pieces of information are important in solving ambiguities?
o Two pieces of information important in solving ambiguities
The new event itself
Knowledge of prior information that can be related and used to interpret the new event
What are requirements for the information processing system?
• Information processing system takes some form of input and turn it into a useful output
What is the information processing system for the brain?
• For the brain: External inputs (sensory information)-> storage and algorithms (memories)-> external outputs
What is a representation? How is its usefulness determined?
o Representation- stored information (in a particular format)
Formal system for making explicit certain entities or types of information, together with a specification of how the system does this
Any particular representation makes certain information explicit at the expense of information that is pushed into the background and may be hard to recover
The usefulness of a representation depends upon how well suited it is to the purpose for which it is used
What is a process?
o Process- an algorithm that changes the form of the representation
Describe Shepard and Metzler (1971)’s representation experiment and its conclusion
o Shepard and Metzler (1971)
Showed subjects either original object or rotated object
The more rotated away from its original orientation the object was, the harder it was to recognise the object as a copy of the original
Suggests that brain is physically rotating the object into the same orientation as the other one to make the comparison
What are Marr’s (1982) levels of explanation?
• The three levels at which any machine carrying out an information processing task must be understood, but as these concepts are somewhat loosely related, some phenomena may be explained at only one or two of the
o Computational theory
o Representation and algorithm (most important level of thinking)
o Hardware implementation
Describe Marr’s computational theory, its important features and examples of its accomplishments.
o Computational theory What is the goal of the computation, why is it appropriate, and what is the logic of the strategy by which it can be carried out? Important features • It contains separate arguments about what is computed and why • Resulting operation is defined uniquely by the constraints it has to satisfy What and why What are we trying to accomplish? • Object recognition • Emotion perception • Store memories • Guided action • Etc.
