NEURO217 Methods

perturbing function study methods

clinical-pathological correlations in neurological/psychiatric cases

pharmacologic interventions

invasive brain stimulation

non-invasive brain stimulation

in silicon lesion using computational modeling/neural networks

human lesion studies advantages

causal evidence

determines essential functions of brain areas

determines what rest of brain can do in absence of it (recovery)

corroborates some cognitive functions in non-human animals by comparing the impact of brain lesions across species

different behavioral consequences emerge with a Brian insult → insights into the kinds of functions the brain area subserves (loss/decreased, release/increased, disorganized/fragmented function)

human lesion studies disadvantages

organic lesions generally do not follow functional or anatomical boundaries

lesion location and extent varies across individual patients and etiologies

reorganization: brain compensates after damage, learning-related plasticity

small sample sizes in most patient studies

lack of baseline assessments before the insult/disease (IQ, personality)

co-morbidities

functional heterogeneity in where cognitive functions reside across individuals

agnosognosia (lack of awareness of behavioral problems)

diaschisis (brain insult in one region can impact function in another region because the regions become disconnected/don’t communicate in a typical manner)

emphasis is on single area of the brain, although cognitive functions tend to be distributed across networks of the brain

lesion-symptom mapping

partial solution to lesion overlap problem

analytical method that associates cognitive functions with specific brain regions by showing common deficits across individuals with similar lesion overlap

3D MRI reconstruction and co-registration methods can extract anatomic locus overlap across individual patients

double dissociation

similar dissociations can be found using neuromonitoring methods but these are correlative, not causative; double dissociation demonstrates that one region is necessary for one cognitive function but not another, and vice-versa for another regionbr

pharmacological intervention

administration of agonists or antagonists that impact neurotransmission

advantages of pharmacological intervention

neurochemical specificity

cross-species comparisons

can compare different doses to get a dose-response curve that characterizes the drug’s impact on the cognitive function of interest

mechanism of action at the synapse can be specific

pharmacological intervention disadvantages

difficulty localizing effects to specific brain regions bc of the large distribution of neurotransmitters in the brain

side effects from drugs

interactions with other neurochemical system

limited temporal precision

blood-brain barrier prevents larger molecules from entering the brain

can have both central and peripheral effects

build up of tolerance

intercranial brain stimulation

implanted deep probe with 6 electrode strips in the temporal lobe, superficial electrode mesh on the cortical surface (electrocorticography eCog)

intracranial brain stimulation advantages

highest spatial localization (mm) of all neuroscience tools in humans

direct measure of electrical activity of neurons

high temporal resolution (ms)

converging evidence with non-human animal electrical recordings

intracranial brain simulation disadvantages

invasive

only used in clinical cases (hist. only epilepsy, but now has developed to treat neurological and psychiatric disease)

small sample size

lack of control over location in epilepsy

Transcranial magnetic stimulation (TMS)

special coil is placed on scalp to induce rapidly-changing magnetic fields that penetrate the cortex a few cm and generate weak electrical currents that depolarize neurons

repetitive TMS

effects can be locally excitatory or inhibitory via long-term potentiation or long-term depression-like mechanisms, respectively, effects on neural excitability cumulate beyond the period of stimulation

rTMS (repetitive TMS)

high frequency (~10-20hz) for LTP-like effects, low frequency (~1-5 Hz), for LTD-like effects, typically 30-45 min duration

Theta bursts (repetitive TMS)

intermittent theta burst stimulation (iTBS) of 3 pulses at 50Hz repeated every 200ms, 2s trains w 8s pauses for LTP-like effects, continuous theta burst stimulation (cTBS) of 3 pulses at 50 Hz every 200ms continuously for LTD-like effects (typically 600 pulses/3min)

TMS single pulse

one current pulse at a time, useful for combining with a neuromonitoring method e.g., fMRI to measure brain responses to current delivery for research

in silico lesion using computational modeling/neural networks

neural netowrk models that come from group-average/individual subject neuroimaging data, can estimate hypothetical dependencies of cognitive functions w/ specific components of the network; done by selectively lesioning (removing) each node of the network sequentially and then seeing effect on cognitive performance

can also estimate optimal location for brain stimulation targeting

advantages of in silicon lesion

cheaper than conducting human studies, safe, can save time/effort in subsequent human invetigations by suggesting best targets for interventionsdi

disadvantages for in silicon lesion

computationally demanding, especially for biologically-plausible neural network models

don’t know how it will translate to humans

don’t know which computational model to choose or which one will work best

need multimodal neuroimaging data from humans to create realistic models

intracranial electrophysiological recording

central part of non-human animal research

can both stimulate and monitor brain activity

usually records action potentials (spikes) but also dendritic field potentials; can be a single-unit (1 neuron) or multi-unit recordings

has the best temporal resolution of all neuroscience techniques but misses the big picture because recording only focuses on a small number of neurons in one area

intracranial electrophysiological recording advantage

population code can be extracted from multi-unit recordings, which gives a measure of average of what neurons are active in that area → provides information into how neural activity Is transformed into behavior

scalp encephalography (EEG)/event-related potentials (ERPs)

commonly used to study sleep stages, but now used to study cognitive processes too

electrical charge goes down the dendrites → voltage gradient lines up perpendicular to the cortical surface and volume conducted to the scalp

all neurons are generating charges which are propagated to the scalp — EEG records the electrical currents

unlike single-unit recordings, does not measure action potentials from axons, but measures activity coming from dendritic fields

what does ongoing EEG reflect?

Oscillations of synchronous firing of millions of neurons which are then summed and amplified over large expanses of the brain

most activity in cognitive processing is in the Gamma range (30-70Hz), some are also in the theta (4-8 Hz); theta can trigger gamma

can determine power at certain frequency: relative power in each band is related to stages of sleep/arousal/alertness and cognitive functions

derivation of ERPs

event-related signal extraction: time-locked averaging

repeated specific stimuli, averaged over many trials

time-locked signal is small and noisy

advantages of EEG/ERPs

direct electrical signal

excellent temporal resolution (tens of milliseconds)

non-invasive

newer analytic methods examine synchronization of EEGs across people

disadvantages of EEG/ERPs

signal is noisy and susceptible to environmental distortions (e.g., overhead lights, external sounds)

need to average over many trials to overcome this noise

deep brain structures do not usually volume-conduct to scalp, best for cortical activity with strong laminar organization

misses neurons parallel to the scalp and subcortical nuclei that have closed electric fields

poor confidence in spatial location, especially for broad ERPs and EEG

requires a discrete signal for time-locked averaging to derive ERPs

thicker hair types = problem acquiring good signal

scalp magnetoencephalography (MEG)

measures complementary magnetic component of electrical fields of neurons at the scalp — good for neurons parallel to the scalp, better for neurons embedded in the sulci rather than gyri

tends to be more expensive than EEG, but strengths and weaknesses are similar

Positron emission tomography (PET) imaging

first whole-brain imaging system

involves infection of radioactive isotope of a substance that can cross the blood barrier to the brain, and positron is emitted when radioactivity begins to decay, which interacts with a local electron and emits gamma rays in opposite directions (detected by cameras located in a ring around the head) = 3-D image

some isotopes can be used to indirectly measure blood flow in the brain or glucose consumption

others measure the distribution of specific receptor molecules that are important for certain diseases

advantages of PET

images uptake of specific neurotransmitters/proteins in the brain

can image long-lasting processes that encompass half-life of radioisotope

non-invasive

disadvantages to PET

expensive (~$2500 per subject)

exposure to radiation

needs a nearby cyclotron to synthesize radioisotopes because of their fast half-life

only 1 image per injection, cannot re-expose people to radiation frequently → smaller sample sizes

spatial resolution worse than fMRI but better than EEG, temporal resolutional is typically worse than with fMRI, EEG, MEG because of the long temporal window for measuring radioactive decay

functional MRI (fMRI)

time-varying images where oxygen is being consumed while participants perform a task

local oxygen depletion → rush of blood flow → change is detected by high-field magnets around the head

blood-oxygenation level dependent (BOLD) signal measures local changes in relaxation of magnetic fields back to standard normal states (hemodynamic) — delayed signal

structural MRI

Gives a static image of anatomy, useful for extracting volumes of brain regions, identifying white matter tracts

advantages of fMRI

can visualize activity of the entire brain at once + subcortical structures

better spatial resolution

cheaper than PET but more expensive than EEG/MEG

in most hospitals, newer high-strength magnetic fields can image even higher spatial resolution

can scan both resting-state and task-based activity

can record EEG, psychophysiological responses, or conduct TMS at the same scanner

disadvantages of fMRI

use same time-locking averaging as with ERPs

noisy, claustrophobic environment, no magnetic materials can be present

temporal resolution not as good as EEG/MEG because signal is sampled every 2-3 s

need to co-register the functional data to the structural MRI data then onto standardized brain space to average across subjects = distortionsfun

functional near-infrared spectroscopy (fNIRS)

laser emitters and light detectors presented on the same side of head quantify light absorption and scattering while light is being fed through and picked up (optical imaging)

external features (bone, skin, etc) are mostly transparent to NIR light

BOLD-related signal where hemoglobin absorbs near IR with different spectra for deoxygenated and oxygen forms

strengths of fNIRS

inexpensive compared to fMRI

portable

less motion restriction

more suitable for use in childrenw

weaknesses of fNIRS

only sensitive to cortical surface

noisy signal

less precise anatomical registration and resolution

gold standard of neuromonitoring approaches

derive brain-behavior correlations that combine information across different methods because each have their own strengths and weaknesses and spatiotemporal/neurobiological sensitivities

always need to consider a study’s findings in the context of the methods used and their relative strengths and weaknesses