Hello all,

We are trying to simulate a DCM data using the same connectivity model presented in DCM paper [1]. We use the spm_dcm_create.m function (SPM5), considering a SNR of 7 and the following parameters:

Load factors.mat (it is a file containing the vectors phot, mot and att, which presents information about the onsets of the three applied conditions; it can be obtained in the SPM web page http://www.fil.ion.ucl.ac.uk/spm/data/attention/)

Enter number of regions [4]

Name for Region 1 [V1]

Name for Region 2 [V5]

Name for Region 3 [PPC]

Name for Region 4 [PFC]

Spm_dcm_create: Basic parameters…

Interscan interval {secs} [3.22]

scans per session e.g. 256  [ 360]

Number of conditions/trials [3]

Name for condition/trial 1? [Photic]

vector of onsets - Photic [phot]

duration[s] (events=0) [10]

Parametric modulation [None]

Name for condition/trial 2? [Motion]

vector of onsets - Motion [mot]

duration[s] (events=0) [10]

Parametric modulation [None]

Name for condition/trial 3? [Attention]

vector of onsets - Attention [att]

duration[s] (events=0) [10]

Parametric modulation [None]

Specify intrinsic connections

-1    1    0    0

1   -1    1    0

0    1   -1    1

0    0    1   -1

Effects of Photic           on regions…and connections

V1       1                                  0 0 0 0

V5        0                                  0 0 0 0

PPC     0                                  0 0 0 0

PFC      0                                  0 0 0 0

Effects of Motion          on regions…and connections

V1       0                                  0 0 0 0

V5        0                                  1 0 0 0

PPC     0                                  0 0 0 0

PFC      0                                  0 0 0 0

Effects of Attention       on regions…and connections

V1        0                                  0 0 0 0

V5        0                                  0 0 1 0

PPC     0                                  0 0 0 1

PFC      0                                  0 0 0 0

Then we obtained the following error:

“Modelled system is potentially unstable: Lyapunov exponent of combined connectivity matrix is 0.61803”

Check the output to ensure that values are in a normal range.”

Best regards,

1-Friston, K.J., Harrison, L. and Penny W. Dynamic causal modeling. NeuroImage 19 (2003) 1273-1302.(please, see figure 24 for more details)

Mayrim Vega Hernández
Neurostatistica Department
Neuroinformatics Subdirection
Cuban Neuroscience Center
Calle 25 esq. 158 Cubanacan, Playa
Havana, Cuba. POBox : 6412,6414
Phones: (53-7) 2717442, (53-7) 2086375