Neuroscience: Current

Functional techniques: methods and principles

Functional neuroimaging as a subtractive approach

Criticisms/limitations

Progress

fMRI on the rise! 1992 -> 2013: 4 -> 150,000 articles! some say revolutionary

Basis: haemodynamics: blood movements and O2: an indirect measure of neural activity.

PET: Positron Emission Tomography: 70's: radioactive isotope injected to blood & scanner detects radioactive waves

Limited to 30 seconds due to fast decay

Decays quickly = no harm

fMRI: functional magnetic resonance imaging: 90's: oxygenated v deoxygenated blood has different magnetic properties: BOLD response (blood oxygenation level dependent)

less invasive (no injections)

BOTH

high spatial resolution: mm in fMRI (still millions of neurons though!)

inferior temporal resolution: sluggish bloodflow: minutes in PET, seconds in fMRI

based on an assumption that ^ bloodflow = ^ O2 = ^ neural activity

higher temporal and spatial resolution

Subtraction

a 'statistical map' displaying the difference in brain activity between two different brain states e.g. speaking v resting

"Pure Insertion"

cognitive component A can be inserted into a task and not influence/affect component B

e.g. colour discrimination task: A is the process of interest with multiple colours, B is the baseline/control with one colour

in neuroimaging:

In neuroimaging:

Raichle language task: open eyes, look at nouns, read nouns aloud, say verbs to match each noun

each stage displayed increase in activity in different brain regions

e.g. listen to spoken word (A) listen to complex noise (B) = regions detected which are active during speech processing (process of interest)

not subtracting everything/ accounting for everything so may inaccurately attribute brain activity

risky to assume there is no link between different cognitive processes: i.e. extreme localisation approach

Some brain regions become less active as task complexity increases (if not involved) - this cannot be observed using other methods!

Cognitivist/Functionalist view VS cognitive neuroscience view

mind can be studied separately from brain

we understand the mind via information we have about the brain, therefore, knowledge about the brain constrains understanding of the mind & cognition.

Inferences

Henson: when making assumption that there is some systematic mapping of psychological function to brain structure (cognitive neuroscience view), then neuroimaging can comprise a DV alongside behavioural data, to distinguish psychological theories

Types of reasoning

Adding a DV in addition to catalogue of behaviour, we can progress understanding of cognition & the mind

e.g.: L/R hand button pressing, muscle excitability (EMG) and dependent variable: BOLD response (fMRI)

Deductive

Inductive

theory -> hypothesis -> observation -> confirmation

observation -> pattern -> hypothesis -> theory

"function-to-structure deduction"

context-dependent, theory-specific, a single experiment

context-independent, assumptions are stronger, several experiments

"structure-to-function induction"

"Where?" it doesn't matter! it only matters that a significant difference exists i.e. if C1 and C2 produce qualitatively different patterns of brain activity then C1 & C2 differ in at least one function (F)

Assumptions?: 1 assumption: the same psychological function (F) will not create different patterns of brain activity within the experiment

Henson: Remember/know study: recollection/ familiarity word lists.

Hypotheses: 1: "single-process" memory strength - quantitative difference
2: "dual-process" remembering= recollection and familiarity, knowing= familiarity only - qualitative difference

fMRI can be used to distinguish by comparing & looking at the difference between conditions.

"Where?" precise location matters! i.e. if C1 and C2 activate the same brain regions, the same function (F) is assumed to be engaged in both

"Systematicity": the same regions are involved in the same functions in ALL contexts ( all experiments)

Henson: verbal STM task: item/list probes, serial order maintainance Y/N

greater BOLD response in fMRI for List > item... dorsal PFC related to task

this region: implicated in timing tasks e.g. finger tapping Catalan hence conclude timing involved in performing serial order tasks.

"S-to-FI": probabilistic, tells us how cognitive functions are implemented in the brain


"F-to-SD": qualitative differences, allow us to distinguish b/w competing cognitive theories.

focus on mapping functioning "S-t-FI" - how much can localization really tell us about the mind? is this just the 'new phrenology'?

cognitive theories aren't necessarily about specific brain responses or brain activity, hence, it may be difficult to address some cognitive theories by using neuroimaging

Page: cognitive psychology is interested in understanding how the mind works not where the brain works - issue with "S-to-FI"

"F-to-SD" can distinguish between cognitive theories, so long as they have specific predictions

Wixted: the validity of the interpretation lives and dies with the validity of the cognitive theory on which it depends, in functional neuroimaging.

not all questions can be answered

but this is true for all/any method/data type

"F-to-SD": the tonotopic map: auditory cortex areas active with different tones/frequencies: these do not represent cognitive functions! this method tells us more about structure and organisation rather than functions?

"S-to-FI": activation doesn't equal cognitive function, tasks performed together may activate an 'overarching' function, the same region may perform multiple functions, some processes require suppression which may cancel out positive activity, absence of activity does not imply absence of cognitive involvement - due to statistical thresholds.

Modularity/ localist viewpoint

Vetter: auditory info in visual cortex
Meyer: visual info in auditory cortex

fMRI: causality? poor temporal resolution, spatial resolution is good but neurons are far smaller so cant be so accurate, specific neural representations don't tell us where that representation came from- other regions?

"S-to-FI": some areas are specific to certain functions! e.g. fusiform face area (FFA) Gazzaley: use functional neuroimaging to investigate suppression of facial features in memory tasks in adults

Schacter & Addis: fMRI found activation in the same brain region for both remembering the past and imagining the future: impairments in both

Gauthier: info about other objects e.g. cars activate FFA

Tell us more than behavioural models?

MRI technology progressing rapidly, becoming less dependent on subtraction, can study networks and connectivity as well as localised functioning

Owen: awareness in vegetative state patients: brain activity can be used to communicate with those otherwise unresponsive e.g. picture playing tennis if yes, picture reading a book if no

Wager: pain is subjective e.g. the placebo effect: physicality vs cognition... can use activation of brain pain networks as a marker: found that placebo effect lowers both actual and affective pain

Combining methods: when and where? fMRI + EEG, intervention effects on brain networks? fMRI + TMS/tDCS

anatomical maps are improving which can aid understanding of functional networks

MRI expensive! sample szes too small and lack statistical power Button

Replicability crisis!

liberal statistical thresholds susceptible to false positives Bennett i.e. adjusting p values