Dear Anja,

> I have a question concerning group analysis in spm99. In the first step
> of statistics, when I want to create the design matrix it is very hard
> to type all this data in. Is it possible to create the Design matrix in
> an matlab window by myself?  I loaded the mat-file of an existing
> Design matrix, but I don't understand the numbers in the array xX.X. I
> understood that the number of rows is the number of pictures of my
> data, and that the columns are the different conditions and sessions.
> Somehow the numbers seem to code for the onset times of stimulation
> (negative ones for no stimulation and positive ones for stimulation?),
> but what exactly do the floats code for?


It would be difficult to v create a fMRI design matrix by hand (the
routine spm_fMRI_design was built to do this).  You could look at the
batch code written by J-B and his colleagues to help you if you wanted
to pursue a 'by-hand' approach.

The following is the help for spm_fMRI_design which contains the
necessary information on structures.

I hope this helps - Karl



>> help spm_fMRI_design

  Assembles a design matrix for fMRI studies
  FORMAT [xX,Sess] = spm_fMRI_design(nscan,RT)
 
  nscan   - n vector {nscan(n) = number of scans in session n}
  RT      - intercans interval {seconds}
 
  xX            - structure describing design matrix
  xX.X          - design matrix
  xX.dt         - time bin {secs}
  xX.RT         - Repetition time {secs}
  xX.iH         - vector of H partition (condition effects)      indices,
  xX.iC         - vector of C partition (covariates of interest) indices
  xX.iB         - vector of B partition (block effects)          indices
  xX.iG         - vector of G partition (nuisance variables)     indices
  xX.Xnames     - cellstr of effect names corresponding to columns
                  of the design matrix
 
  Sess{s}.BFstr    - basis function description string
  Sess{s}.DSstr    - Design description string
  Sess{s}.rep      - session replication flag
  Sess{s}.row      - scan   indices      for session s
  Sess{s}.col      - effect indices      for session s
  Sess{s}.name{i}  - of ith trial type   for session s
  Sess{s}.ind{i}   - column indices      for ith trial type {within session}
  Sess{s}.bf{i}    - basis functions     for ith trial type
  Sess{s}.sf{i}    - stick functions     for ith trial type
  Sess{s}.ons{i}   - stimuli onset times for ith trial type (secs)
  Sess{s}.pst{i}   - peristimulus times  for ith trial type (secs)
  Sess{s}.Pv{i}    - vector of paramters for ith trial type
  Sess{s}.Pname{i} - name   of paramters for ith trial type
 
  saves SPM_fMRIDesMtx.mat (xX Sess)
 _______________________________________________________________________
 
  spm_fMRI_design allows you to build design matrices with separable
  session-specific partitions.  Each partition may be the same (in which
  case it is only necessary to specify it once) or different.  Responses
  can be either event- or epoch related, where the latter model prolonged
  and possibly time-varying responses to state-related changes in
  experimental conditions.  Event-related response are modelled in terms
  of responses to instantaneous events.  Mathematically they are both
  modelled by convolving a series of delta (or stick) functions,
  indicating the onset of an event or epoch with a set of basis
  functions.  These basis functions can be very simple, like a box car,
  or may model voxel-specific forms of evoked responses with a linear
  combination of several basis functions (e.g.  a Fourier set).  Basis
  functions can be used to plot estimated responses to single events or
  epochs once the parameters (i.e.  basis function coefficients) have
  been estimated.  The importance of basis functions is that they provide
  a graceful transition between simple fixed response models (like the
  box-car) and finite impulse response (FIR) models, where there is one
  basis function for each scan following an event or epoch onset.  The
  nice thing about basis functions, compared to FIR models, is that data
  sampling and stimulus presentation does not have to be sychronized
  thereby allowing a uniform and unbiased sampling of peri-stimulus time.
 
  spm_fMRI_design allows you to combine both event- and epoch-related
  responses in the same model.  You are asked to specify the number
  of condition (epoch) or trial (event) types.  Epoch and event-related 
  responses are modeled in exactly the same way by first specifying their
  onsets [in terms of stimulus onset asynchronies (SOAs) or explicit onset 
  times] and then convolving with appropriate basis functions (short ones
  for event-related models and longer ones for epoch-related respones).
  Enter 0 to skip these if you only want to use regressors you have designed 
  outside spm_fMRI_design).
  
  Interactions or response modulations can enter at two levels.  Firstly
  the stick function itself can be modulated by some parametric variate
  (this can be time or some trial-specific variate like reaction time)
  modeling the interaction between the trial and the variate or, secondly
  interactions among the trials themselves can be modeled using a Volterra
  series formulation that accommodates interactions over time (and therefore
  within and between trial types).  The first sort of interaction is
  specified by extra (modulated) stick functions in Sess{s}.sf{i}.  If
  a polynomial expansion of the specified variate is requested there will
  be more than one additional column.  The corresponding name of the
  explanatory variables in X.Xname is Sn(s) trial i(p)[q] for the qth
  order expansion of the variate convolved with the pth basis function
  of the ith trial in the sth session.  If no parametric variate is
  specified the name is simply Sn(s) trial i(p).  Interactions among
  and within trials enter as new trial types but do not have .pst or .ons
  fields.  These interactions can be characterized later, in results, in
  terms of the corresponding second order Volterra Kernels.
 
  The design matrix is assembled on a much finer time scale (X.dt) than the 
  TR and is then subsampled at the acquisition times.
 
  Sess{s}.ons{i} contains stimulus onset times in seconds relative to the
  timing of the first scan and are provided to contruct stimuli when using 
  stochastic designs
 
 
  Notes on spm_get_ons, spm_get_bf and spm_Volterra are included below
  for convenience.
 
                            ----------------
 
  spm_get_ons contructs a cell of sparse delta functions specifying the
  onset of events or epochs (or both). These are convolved with a basis set
  at a later stage to give regressors that enter into the design matrix.
  Interactions of evoked responses with some parameter (time or a specified 
  variate Pv) enter at this stage as additional columns in sf with each delta
  function multiplied by the [expansion of the] trial-specific parameter.
  If parametric modulation is modeled, P contains the original variate and
  Pname is its name.  Otherwise P{i} = [] and Pname{i} = '';
 
                            ----------------
 
  spm_get_bf prompts for basis functions to model event or epoch-related
  responses.  The basis functions returned are unitary and orthonormal
  when defined as a function of peri-stimulus time in time-bins.
  It is at this point that the distinction between event and epoch-related 
  responses enters.
 
                            ----------------
 
  For first order expansions spm_Volterra simply convolves the causes
  (e.g. stick functions) in SF by the basis functions in BF to create
  a design matrix X.  For second order expansions new entries appear
  in IND, BF and name that correspond to the interaction among the
  orginal causes (if the events are sufficiently close in time).
  The basis functions for these are two dimensional and are used to
  assemble the second order kernel in spm_graph.m.  Second order effects
  are computed for only the first column of SF.
 


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%