Experimental design

Table 6: Experimental designs in functional neuroimaging.
Type Subtype Description Reference
Subtraction     [Lassen et al., 1978]
Double subtraction fMRI Subtraction of the form $ (A-B)-(C-B)$ [Poldrack et al., 1998,Poldrack et al., 1999]
Parametric PET   [Grafton et al., 1992]
Parametric   Presentation rate of heard words is varied [Price et al., 1992]
Parametric Nonlinear   [Büchel et al., 1996]
Parametric     [Büchel et al., 1998]
Block   Block-design in fMRI presents the paradigm  
Event-related   fMRI design where the paradigm is presented as short stimuli.  
Mixed   fMRI design Mixed between block design and event-related. e.g., [Visscher et al., 2003,Donaldson, 2004]
Stochastic event-related   Event-related design where the periods between the stimuli are varied.  
$ 2\times 2$ factorial PET Motor activation and time [Friston et al., 1992a]
$ 2\times 2$ factorial PET Memory and drug (apomorphine and buspirone) interaction [Friston et al., 1992b]
$ 2\times 2$ factorial efMRI Maintenance and manipulation [Honey et al., 2001]
PPI   Pycho-physiological interaction where the change in interaction between one area and another is analyzed [Friston et al., 1997], Example analysis:
Conjunction   $ (A-B) \wedge (C-D)$ [Price and Friston, 1997,Friston et al., 1999a,Nichols et al., 2005,Brett et al., 2004,Friston et al., 2005]
Conjunction Multimodal   [Hayasaka et al., 2006]

Table 6 shows some of the experimental designs in (cognitive) functional neuroimaging.

In pycho-physiological interaction (PPI) experiments [Friston et al., 1997] the change in interaction strength between one area and another are investigated. The change appears when the experimental paradigm is changed.

Cognitive conjunction [Price and Friston, 1997] includes a series of subtraction designs that differ in the cognitive component they elicit. The resulting statistical maps are ``and'ed'' to form a statistical map for (usually) a single cognitive component. The exact interpretation of the ``and'' can vary depending on the statistics used [Brett et al., 2004].

(Hemodynamic) response modeling in fMRI can model shape and amplitude of hemodynamic response function (HRF), e.g., [Boynton et al., 1996] where the function between the contrast of a visual stimuli (flickering checkerboard pattern) and the BOLD response is described.

In fMRI and especially in event-related fMRI (efMRI) the paradigm is either a ``Dirac'' impulse or a short block of stimulus/response. Common for efMRI is that the measured signal is not regarded as being in a steady state. In these kind of experiments the paradigm pattern can be varied in many ways, e.g., the interstimulus interval (ISI) can be varied. Jittering is when the presentation of the paradigm is deliberately varied so that it is not correlated in phase with the acquisition rate of the scanner. In stochastic designs the paradigm is presented aperiodically [Heid et al., 1997].

A program for optimizing experiment design with genetic algorithms has been constructed by Tor Wagner and Tom E. Nichols [Wager and Nichols, 2002]:

Statistical approach to optimize experimental design

[Liu et al., 2001a] distinguishes between estimation efficiency and detection power, where the former is the ability to estimate the shape of the hemodynamic response function (HRF) and the latter is the ability to estimated the brain activation. In connection with the general linear model the measure for the efficiency is [Dale, 1999, equation 12], [Friston et al., 1999b, page 608 and equation A.3], [Liu et al., 2001a, equation 4], [Birn et al., 2002, equation 12]

$\displaystyle E = \frac{1}{\mbox{trace} \left[ ({\bf X}^{\sf T}{\bf X})^{-1} \right] }$ (1)

where $ {\bf X}$ is the design matrix and $ \sigma^2$ is the covariance of the Gaussian distributed noise. This is called the inverse A-optimality criterion [Montgomery, 2001, page 468].

The direct optimization of experimental design for BOLD fMRI should take into account the nonlinear effect occurring when a stimulus is presented rapidly and the 1/f-noise which prevents the stimulus to be presented to long. [Aguirre et al., 2002] find that BOLD fMRI are contaminated by autocorrelated noise contrary to perfusion arterial spin labeling (ASL) fMRI in which the autocorrelated noise is absent.

Other discussions of functional neuroimaging experimental design are found in [Birn et al., 2002].


J. E. Desmond, G. H. Glover, Estimating sample size in functional MRI (fMRI) neuroimaging studies: statistical power analyses. J Neurosci Methods 2002 Aug 30;118(2):115-28 PMID: 12204303

Mapping Voxel-Based Statistical Power on Parametric Images" by JD Van Horn et al. Neuroimage 7, 97-107 (1998)

Jason Steffener SPM-mailing list

Neuropsychological tests

Table 7 shows a number of tests that are particularly used in clinical neuropsychology. Description of many neuropsychological tests are available in [Crawford et al., 1992,Lezak, 1995].

Table 7: Neuropsychological tests.
Abbrev. Name Description Reference
ADIS Adult Diagnostic Interview Schedule    
BPRS Brief Psychiatric Rating Scale   Southwick et al., 1993
CRAS Clinician Rated Anxiety Scale    
EPQ Eysenck Personality Questionnaire    
HAMA Hamilton Anxiety Rating Scale ``Hamilton Anxiety scale'' [Hamilton, 1959]
HAMD Hamilton depression scale    
MAS Manifest Anxiety Scale   [Taylor, 1953]
MMPI Minnesota Multiphasic Personality Inventory   Wikipedia
MMSE Mini-Mental State Examination A brief test for dementia [Folstein et al., 1975], [Morris and Kopelman, 1992, page 304]
NEO PI-R Revised NEO Personality Inventory   [Costa and McCrae, 1992]
PANAS Positive Affect Negatve Affect Schedule   [Watson et al., 1988]
PASS Panic Attack Symptom Scale A panic attack is defined in DSMIV as a PASS score of higher than 8  
POMS Profile of Mood States. Originally ``Psychiatric Outpatient Mood Scale''   [McNair et al., 1981,McNair and Lorr, 1964]
SAM Self-Assessment Manikin A quick non-verbal pictorial assessment of emotional reaction split in pleasure, arousal, and dominance components [Bradley and Lang, 1994]
SCID Structured Clinical Interview for DSM-IV   [Spitzer et al., 1987]
SCL-90-R Symptom Check List 90 Revised   Hopkins(?)
SHSS Stanford Hypnotic Susceptibility Scale ``Form A'' ``Form C'', ``Form I'' and Form II'' seem to exist [Hilgard et al., 1963]
SDS Self-rating Depression Scale. Also ``Zung's depression scale''. 20 questions answered by the patient. $ <50$ is normal, $ >70$ is severe depression [Zung et al., 1965]
STAIS-s Spielberger's State Anxiety Inventory    
SUB Subjective Units of Distress    
UPDRS Unified Parkinson's Disease Rating Scale   [Fahn et al., 1987]
WAIS Wechsler Adult Intellingence Scale   [Wechsler, 1955,Crawford, 1992]
WAIS-R Wechsler Adult Intellingence Scale -- Revised   [Wechsler, 1981,Crawford, 1992]
Y-BOCS Yale-Brown obsessive-compulsive scale   [Goodman et al., 1989]
YGTSS Yale Global Tic Severity Scale Used for Tourette Syndrome

Psychological test software tools

Table 8 displays a list of psychological experiment generator software. Some of these only incorporate presentation of visual stimulus, while others also allow for collection of responses from a number of different devices as well as synchronization with an fMRI scanner. Psychology Software Distribution,, maintains a short list of experiments written for a number of these software packages. It is based on data collected by CTI from University of York, [Hammond and Trapp, 1996] is a 1996 review of ten experiment software packages. Journal of Neurobehavioral Experiments and Stimuli (JONES, also lists a experiments.

Table 8: Psychological software and hardware.
Name   Description Reference
BrainLogics A(B) Commercial fMRI system with response buttons, experimental development and analysis
Cogent A Windows Matlab-based presentation program with possibility for manipulating output (sound and graphics) and input (mouse, keyboard, joystick, serial and parallel port plus fMRI scanner triggering) [Hutton et al., 2002]
DMDX A   [Forster and Forster, 2003]
Eloquence A(B) Commercial paradigm presentation hardware and software for fMRI
E-Prime A Commercial Windows-based presentation program with graphical and Visual Basic like scripting language and response collection. From Psychology Software Tools.
ERTS A PC-based presentation program. Features fMRI triggering with a TTL-signal
FEST A ``FMRIB Enhanced Stimulation Tool'': Stimulation tool for presenting pictures, sounds and text and record button press.
IFIS A(B) Integrated Functional Imaging System. Commercial software and hardware system. Software is a variation of E-Prime suitable for presentation in an functional neuroimaging experiment. Features MRI RF-pulse triggering (via antenna), graphical and Visual Basic like scripting language, non-magnetic finger buttons and dual screens (control and presentation). fMRI analysis via BrainVoyager. IFIS-SA:
LabVIEW A Commercial program from National Instruments. Signal aquisition program and stimulus control, SPM mailing list: labview
MacStim A 68k Mac based program by David Darby
MEL A DOS-based experiment generation program.
Presentation A PC-based stimulus generator with input devices (e.g., mouse and joystick) and MRI-scanner synchronization with built-in programming language
Psychophysics Toolbox A Matlab-based (Windows and MacIntosh) toolbox for vision research., [Brainard, 1997,Pelli, 1997]
PsyScope A Macintosh-based (OS7-OS9) ``interactive graphic system for experimental design and control''. Development has ceased. [Cohen et al., 1993,MacWhinney et al., 1997], (original site), (Belgian mirror)
PyEPL A ``The Python Experiment-Programming Library'': Python-based with playback and recording of sound, displaying text, images and 3 dimensional environments with input from keyboard and synchronization with external recorded events [Geller, 2006,Geller et al., 2006],
Stim$ ^2$ A Commerical Windows-based stimulus presentation from Neuroscan.
Rings A Mac-based program for generation of simple visual stimuli
SuperLab A Commercial Macintosh/Windows-based experiment generator supporting a number of input devices (keyboard, mouse, microphone, Cedrus and PST response boxes, I/O cards). From Cedrus Corporation.
Vision Egg A 2D and 3D visual stimulus creation and control open source software based on python and OpenGL for Windows, Mac OS X, Linux, SGI

MRI compatible hardware

Table 9 shows publicly available MRI compatible hardware. A fiber optic joystick is used in [Van Horn et al., 2001b]. The Altra Felix pointing device is not made for use in MRI, but it has been used in a 1.5T scanner [Balslev et al., 2004].

Table 9: MRI compatible hardware: Response pads, joysticks, displays, earphones and EEG recording.
Name Items Reference
ADInstruments Finger electrodes, galvanic skin response, respiratory transducer, laser Doppler probe, termocouple probe
ASL 504LRO fMRI compatible eyetracker
Avotec LCD projection system with eye tracking, e.g., ``Silent Vision'' system.
Cambridge Research Systems ``Lumina'' response pads and ``MR-Eyetracker'' E.g.,
Coldswitch LUMItouchTM response pad and LUMItouch Joystick
Current Designs Fiber Optic Response Pad system (fORP): Buttons, joystick, trackball
Imagelys fMRI response pads and fMRI synchronization interface
Invivo Research Patient monitoring, Pulse Oximetry, and e.g., ``monitors ECG, Respiration, HR, SpO2, NIBP, with optional EtCO2''
Mag Design and Engineering Response box, head constraints, tactile stimuli, eye tracking, joystick. Commercial products by Ben Krasnow
Measurand ``ShapeHand MRI'' data glove
MRI Devices Corporation IFIS, response pad, earphones, visual display, control room console
NeuroScan STIM stereoscopic visual (Silent VisionTM) and stereo auditory (Silent ScanTM) stimulation, MagLink EEG recording
Nonin Portable Pulse Oximeter>
Resonance Technology Commander XG earphones, MRIVision 2000 and VisuaStim XGA stereoscopic head-mounted displays
Rowland Institute Response box by Chris Stokes(?)
Sven Haller Pneumatic MR compatible response box available ``at a reasonable price'' SPM Mailing list, 2002-9-11

Finn Årup Nielsen 2010-04-23