-
Notifications
You must be signed in to change notification settings - Fork 2
/
spm.man
291 lines (288 loc) · 13.6 KB
/
spm.man
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
% Statistical Parametric Mapping - SPM5
% ______________________________________________________________________
%
% ___ ____ __ __
% / __)( _ \( \/ )
% \__ \ )___/ ) ( Statistical Parametric Mapping
% (___/(__) (_/\/\_) SPM - http://www.fil.ion.ucl.ac.uk/spm/
%
% ______________________________________________________________________
%
% Statistical Parametric Mapping refers to the construction and
% assessment of spatially extended statistical process used to test
% hypotheses about [neuro]imaging data from SPECT/PET & fMRI. These
% ideas have been instantiated in software that is called SPM.
% This software also deals with other issues in image analysis
% such as spatial registration and normalisation problems.
%
% ----------------
%
% Please refer to this version as "SPM5" in papers and communications.
%
% ----------------
% Contents:
% 1) SPM - The software
% 2) Release notes for SPM5
%
%
% ======================================================================
% 1) S P M - T h e s o f t w a r e
% ======================================================================
%
% SPM was written to organise and interpret our data (at the Wellcome
% Department of Cognitive Neurology, and previously at the MRC
% Cyclotron Unit, London UK). The distributed version is the same as
% that we use ourselves.
%
% SPM is made freely available to the [neuro]imaging community, to
% promote collaboration and a common analysis scheme across
% laboratories.
%
% ______________________________________________________________________
% Authors
%
% SPM is developed under the auspices of The Wellcome Department of
% Imaging Neuroscience, a department of the Institute of Neurology at
% University College London.
%
% SPM94 was written primarily by Karl Friston in the first half of
% 1994, with assistance from John Ashburner (MRC-CU), Jon Heather
% (WDoIN), and Andrew Holmes (Department of Statistics, University of
% Glasgow). Subsequent development, under the direction of Prof. Karl
% Friston at the Wellcome Department of Imaging Neuroscience, has
% benefited from substantial input (technical and theoretical) from:
% John Ashburner (WDoIN), Andrew Holmes (WDoIN & Robertson Centre for
% Biostatistics, University of Glasgow, Scotland), Jean-Baptiste Poline
% (WDoIN & CEA/DRM/SHFJ, Orsay, France), Christian Buechel (WDoIN),
% Matthew Brett (MRC-CBU, Cambridge, England), Chloe Hutton (WDoIN) and
% Keith Worsley (Department of Statistics, McGill University, Montreal,
% Canada).
%
% SPM5 was developed by: Jesper Andersson (unwarping), John Ashburner
% (spatial, image handling), Karl Friston (project director,
% fMRI design), Andrew Holmes (statistics, user interfaces),
% Jean-Baptiste Poline (statistics), Philippe Ciuciu, Tom Nichols
% (statistics), Matthew Brett (WinNT implementation and other stuff),
% Volkmar Glauche (visualisation) & Darren Gitelman (various bits).
% Theoretical input was provided by Christian Buchel, Daniel Glaser,
% Rik Henson, Stefan Kiebel, Will Penny, Keith Worsley and loads of
% other people. We would like to thank everyone who has provided
% feedback on SPM.
%
% We envisage that this software will be used in a diverse number of
% ways. Although SPM has grown out of a PET background, it is now an
% established package for the analysis of fMRI data, and is also used
% for structural data.
%
% ______________________________________________________________________
% Resources
%
% The SPMweb site is the central repository for SPM resources:
% http://www.fil.ion.ucl.ac.uk/spm/
% Introductory material, installation details, documentation, course
% details and patches are published on the site.
%
% There is an SPM eMail discussion list, hosted at
% <[email protected]>. The list is monitored by the authors, and
% discusses theoretical, methodological and practical issues of
% Statistical Parametric Mapping and SPM. Subscribe by sending the one
% line message: "join spm firstname lastname" to
% <[email protected]>. (Users at NIH or UC-Davis should join
% their local SPM feeds.) The SPMweb site has further details:
% http://www.fil.ion.ucl.ac.uk/spm/support/
%
% ----------------
%
% In order to use the advanced spatial, statistical modelling and
% inference tools of SPM, it is vital to have at least a conceptual
% understanding of the theoretical underpinnings. Therefore, we
% recommend the theoretical articles published in the peer reviewed
% literature, and the SPMcourse notes (available from the SPMweb site).
%
% ----------------
%
% Please report bugs to the authors at <[email protected]>
% Peculiarities may actually be features(!), and should be raised on the
% SPM eMail discussion list, <[email protected]>.
% ______________________________________________________________________
% The SPM distribution
%
% The SPM software is a suite of MATLAB functions, scripts and data
% files, with some externally compiled C routines, implementing
% Statistical Parametric Mapping. MATLAB, a commercial engineering
% mathematics package, is required to use SPM. MATLAB is produced by The
% MathWorks, Inc. Natick, MA, USA. http://www.mathworks.com/
% eMail:[email protected]. SPM requires only core MATLAB to run (no
% special toolboxes are required).
%
% SPM5 is written for MATLAB versions 6.5.1 to 7.2 under UNIX, LINUX
% and Windows. SPM5 may not will not work with versions of MATLAB prior
% to 6.5. Later versions of MATLAB (released after SPM5), will probably
% need additional patches in order to run. Once developed, these will
% be made available from:
% ftp://ftp.fil.ion.ucl.ac.uk/spm/spm5_updates/
%
% Binaries of the external C-mex routines are provided for Solaris2,
% Linux and Windows only. Users of other platforms need an ANSI C
% compiler to compile the supplied C source (Makefile provided).
% ______________________________________________________________________
% Copyright & licencing
%
% SPM (being the collection of files given in the manifest in the
% Contents.m file) is free but copyright software, distributed under
% the terms of the GNU General Public Licence as published by the Free
% Software Foundation (either version 2, as given in file
% spm_LICENCE.man, or at your option, any later version). Further
% details on "copyleft" can be found at http://www.gnu.org/copyleft/.
%
% SPM is supplied as is.
% No formal support or maintenance is provided or implied.
% ______________________________________________________________________
% File formats
%
% The various file types included in SPM are:
%
% spm_*.m files: ASCII files that form the main structure for SPM.
% Most of SPM is written as MatLab functions. These are compiled
% at their first invocation, leading to a slight delay in the
% startup of some routines. MatLab script files are occasionally
% used. These are interpreted by MATLAB, but have the advantage
% of working in the base MatLab workspace, such that their
% results are available to the user after completion.
%
% Clearly MatLab is slower than writing everything in fully optimised
% C; however the fundamental advantage of having a didactic pseudo-code
% specification of this sort is preferred over implementational
% efficacy. Further, MatLab *is* optimised for matrix and vector
% operations, which are utilised whenever possible.
%
% src/spm_*.c: ASCII files that are complied in a MATLAB-specific
% fashion to produce programs that can be called directly from
% MATLAB. Once compiled these routines are suffixed in a
% platform dependent fashion (e.g. spm_*.mexsol or mexglx). These
% routines implement memory mapping and some numerical and image
% operations that are called frequently. Precompiled Mex files
% are provided for Solaris2, Linux and Windows platforms, a
% Makefile is included for other platforms.
%
% spm_*.man: ASCII files containing manual pages.
%
% *.mat MATLAB specific data files that can be loaded directly
% into MATLAB. These files contain images and other data in matrix
% format, usually in double precision (see MATLAB user's guide)
%
% Where possible the user interface and computational or analytical
% aspects of the software have been segregated such that spm_*_ui.m
% sets up the user interface and assembles the appropriate input
% arguments for spm_*.m. spm_*.m contains the statistical and
% mathematical implementation of a generic nature. spm_*.m would be of
% greater interest to those whose wish to incorporate SPM into an
% existing package with its own 'front end'.
%
% ======================================================================
% 2) Release Notes for SPM5
% ======================================================================
%
% EEG/MEG analysis
%
% A major new feature of SPM5 is the functionality to analyse epoched
% EEG and MEG data. This functionality can be divided into two
% components: (i) preprocessing and (ii) statistical analysis.
% Preprocessing contains the usual steps; e.g. epoching, filtering,
% artifact detection, time-frequency decomposition and averaging. After
% preprocessing, the data is projected into voxel space. This space can
% be either the 2D-scalp (interpolation) or 3D-brain space (source
% density reconstruction - see below). Conversion to voxel-space allows
% one to use SPM's existing functionality for model specification,
% parameter estimation and adjusting p-values. In particular, we make
% use of Random Field Theory to adjust p-values to account for the
% multiple comparisons over brain space.
%
%
% Time-frequency EEG/MEG analysis
%
% This combines the time-frequency components of preprocessing for
% basic EEG/MEG pre-processing with extant Random Field Theory adjusted
% p-values. In this application the adjustment is for searches over 2D
% time-frequency domains. There have been suitable extensions for
% graceful and more general handling of 2D SPMs.
%
%
% Source localisation
%
% Features will include:
% - Co-registration of structural MRI (sMRI) and EEG/MEG
% - Individual forward solutions based on a deformed template mesh of
% the cortical surface. This is based on a simple spherical forward
% solution provided by BrainStorm
% - A forward solution based on the 3 sphere shell model
% - Distributed source localisation using Empirical Bayes with
% smoothness and/or functional priors.
% - Equivalent Current Dipole (ECD) source localisation
% - Interpolation of the reconstructed source activity into sMRI
% voxel-space
%
%
% Spatial normalisation and segmentation
%
% Because of the confusion about "optimized VBM" and "customized
% templates", there will be a new integrated spatial normalisation and
% segmentation routine. This enables spatial normalisation of images
% acquired using a wider range of sequences (so fewer "customized
% templates" are likely to be needed). More accurate results can
% potentially be achieved by combining tissue classification, bias
% correction and nonlinear warping into the same generative of forward
% model.
%
%
% Bayesian fMRI analysis with spatial priors
%
% This analysis method potentially improves the sensitivity of fMRI
% analysis by spatially regularising signal and noise parameters using
% variational inference in the context of a Gaussian Markov Random
% Field. Smoothing fMRI data an arbitrary amount in a pre-processing
% stage is no longer required, as the optimal amount of spatial
% regularisation can be determined automatically and separately for
% each putative experimental effect.
%
%
% User interface
%
% There will be a new User-Interface (UI) for the pre-processing steps.
% Some stats utilities may also have the new UI. The idea is to allow
% easier batching, and flexibility of the order in which operations are
% defined. The batch files should serve as documentation about how the
% data were processed. The user-interface will include explanations of
% what the various options mean. Most of the code behind SPM2 is just
% for the user interface. Code for the SPM5 user-interface should be
% much simpler.
%
%
% Data formats
%
% Data will be written as NIFTI-1 format instead of Analyze, although
% the old Analyze format will still be read directly by the various SPM
% functions. MINC and ECAT7 will no longer be read directly, but will
% instead need to be converted. Tools will be provided for doing this.
% For more information on NIFTI-1, see http://nifti.nimh.nih.gov/dfwg/.
% The motivation for this is to reduce confusion about whether the
% images are flipped or not, and increase inter-operability among
% packages.
%
%
% DCM for ERPs
%
% DCM for ERPs is a toolbox with a standalone GUI that enables
% inferences about large scale neural networks, based on epoched ERP
% data. The user interface allows you to specify the data,
% experimental design and lead fields (that implicitly specify the
% sources or nodes of the network). Having specified this information,
% and the network connectivity, a model is constructed and identified
% in terms of the conditional distribution of the its parameters using
% EM. In addition, the log-evidence, for that particular model, is
% returned to facilitate Bayesian model comparison.
%
%__________________________________________________________________________
% Copyright (C) 2005 Wellcome Department of Imaging Neuroscience
% The FIL Methods Group
% $Id: spm.man 1032 2007-12-20 14:45:55Z john $