PSGY4009

Experimental design for functional imaging

Denis Schluppeck

2024-10-03

1 Today

  • Who am I? Who is everyone?
  • A quick overview of the module.
  • What to expect, assessment/coursework?
  • 101 of functional MRI and experiment design.

2 Overview

Timetable week Lecturer Lecture
02 D Schluppeck Overview, fMRI & study design.
03 M Schürmann Basic neuroanatomy
04 K Dyke Brain stimulation & study design.
05 W v Heuven Language
06 Reading week (!)
07 D Schluppeck Vision + brain imaging
08 D Schluppeck Q&A, experimental design, coursework
09 L Cragg Developmental neuroimaging.
10 J Derrfuss Cognitive control, attention, and working memory.
11 R Filik Moral Cognition
12 D Schluppeck Perceptual Decision making, Q&A for assignment

3 Assessment

  • Written assignment (max 3000 words) including a 250 word abstract.

  • Details on moodle (2024/25).

3.1 What? How?

How can fMRI and/or brain stimulation be used to study different neuroscience questions? Cover at least two topics and/or methods from the course.

4 Today’s lecture

  • A brief introduction to fMRI.

  • Scope and limitations of the fMRI method.

  • Principles of experimental design in psychology.

  • Experimental design for fMRI.

5 Learning objectives

By the end of the lecture you should be able to:

  • Understand what fMRI measures.
  • Evaluate the strengths and weaknesses of fMRI.
  • Evaluate different experimental designs.

image of statistical maps on brains

6 LO

fMRI basics

7 fMRI basics

  • MRI uses static and oscillating electromagnetic fields to detect different biological tissue properties.

  • functional MRI measures changes in blood oxygenation over time.

  • BOLD signal: blood oxygen level dependent

8 Example research questions

Where are the areas more active in Condition A than in Condition B?

How do pre-defined regions of interest respond in different experimental conditions?

9 Experimental “pipeline”

Prepare

Theory ➡️ Research Question ➡️ Hypothesis

Plan & do

Experimental Design ➡️ Piloting ➡️ Data collection

Analyse & report

Preprocessing ➡️ Analysis ➡️ Interpretation

10 Study

11 General thoughts (BOLD)

What does the blood-oxygen-level-dependent (BOLD) signal measure?

  • neuronal activity?!
  • populations of neurons (excitatory, inhibitory)
  • spikes or (pre-)synaptic activity

12 General thoughts (Signal/Noise)

(fMRI) signals are weak + noisy.

  • How does that limit voxel size?
  • Short events, minimum durations?
  • Need for averaging?
  • Blocks versus “event-related” designs?

13 General thoughts (Variability)

Details of brain anatomy vary across individuals.

  • What does this mean for group studies?
  • Individual subjects results first?
  • Data in 3d or 2d (cortical surface?)

14 Example data / images

anatomy

Anatomical image (T1w)

fMRI image

Functional image (T2*, EPI)

15 fMRI measures changes over time

16 Example: visual cortex

anatomical image with position of fMRI slice

17 Example: visual cortex (movie)

one slice over 160 repeated measurements (every 1.5s)

18 fMRI measures changes over time

19 “There is no free lunch”

  • both, spatial resolution and temporal resolution, are limited
  • often, experimenter has to trade off one against the other
  • so, different choices for different applications…

20 Why do we need high spatial resolution?

  • brain structures of interest are ~mm in size, sometimes separated by ~cm; need appropriate sampling

  • smaller voxels: less mixing of grey matter, white matter, CSF, veins, … reduced partial voluming

21 Appropriate sampling

Table of structures / sizes

22 Volume → no. of cells

… if packing density in grey matter of cortex is ~50,000 cells / mm\(^3\)

edge (mm) volume (mm\(^3\)) #cells in pure GM
1 1 50k
3 27 1.35M
5 125 6.25M

23 PVE

24 Partial volume effects

A typical 3mm voxel contains a mixture of tissues

25 What limits (spatial) resolution

hardware (the scanner)

the subjects

  • peripheral nerve stimulation PNS
  • specific absorption rate SAR limits
  • time in scanner ( >1.5h is not fun)

signal-to-noise ratio - smaller voxels = proportionally more noise head motion, physiology, …

26 fMRI measures changes over time

27 What limits (temporal) resolution

physiology

basis of the blood-oxygen-level-dependent (BOLD) signal

  • BOLD is haemodynamic, an indirect measure of neural activity (+ still hotly debated)
  • BOLD signal is delayed and blurred in space, but also in time

28 Haemodynamic response

the shape of the response to a brief event is called the haemodynamic response function (HRF)

haemodynamic response

e.g. haemodynamic response to a 1s visual stimulus peaks several seconds later and is spread out

29 Intermediate summary

  • Overview of the fMRI method.

  • How to interpret fMRI data.

  • Limiting factors of fMRI.

  • Trade-off between temporal and spatial resolution.

Thinking about temporal resolution particularly important when designing fMRI experiments

30 Some principles of experimental design

Why do we need to worry about designing experiments?

31 Terminology

Independent Variable:

variable which is intentionally manipulated (different manipulations are called conditions / levels)

Dependent Variable:

measured variable

Confound:

a property that co-varies with the IV

32 eg. this experiment

33 Dealing with confounds

Counterbalancing: a process for removing confounds by ensuring they have equal influence at each level of the IV

Randomization: removing confounds by making sure they vary randomly at each level of the IV

34 Within or between subjects?

Within-subjects designs are stronger as they control for:

  • Age Gender.
  • Experience.
  • Individual Difference.
  • Brain size / shape.

Between-subjects designs are sometime unavoidable

  • Diagnostic comparisons
  • Gender effects

Match the design to the research question…

35 … and for fMRI?

All the above, plus…

  • Selecting an appropriate IV
  • Selecting an appropriate control condition
  • Types of experimental design
  • Stimulus presentation timing

36 Many studies: cognitive subtraction

Compare two very similar conditions that are assumed to differ only in one property.

37 Control conditions are important

38 fMRI designs

  • Block Designs – trials appear in alternating blocks (Task A, Task B)

  • Event Related Designs – stimuli are presented one-at-a-time in a random order

  • Mixed Designs – alternating blocks of tasks but with mixed trials

  • (resting state fMRI) - no task

39 Block designs

  • Powerful for detecting signal changes if a good control is selected.

  • Not very good for estimating the shape of the HRF.

  • Sensitive to noise (scanner drift)

  • The block length is very important to consider

41

42

43 Summary

Principles of experimental psychology apply to good fMRI study design also.

  • Consider how the hypotheses make predictions about space and time (and do they align with the scope of what fMRI can measure).

  • The experimental design should be appropriate for the research question.

  • Timing of stimulus presentation is key!

44 Thanks

See you soon!

45 Colophon

  • This presentation was made with quarto and revealjs.

  • Uses a font called Atkinson Hyerlegible, which was designed to work better for people with low vision: available via google fonts.

https://schluppeck.github.io/edffi