Differences

This shows you the differences between two versions of the page.

Link to this comparison view

reference:harmony [2018/08/23 14:43] (current)
Line 1: Line 1:
 +=====  HARMONY =====
 +
 +Note that this reference documentation is identical to the help that is displayed in MATLAB when you type "help harmony"​.
 +
 +<​html><​pre>​
 +  <a href=/​reference/​harmony><​font color=green>​HARMONY</​font></​a>​ computes a linear estimate of the current in a distributed source model
 +  using a mesh harmonic based low-pass filter.
 + 
 +  Use as
 +    [dipout] = minimumnormestimate(dip,​ grad, headmodel, dat, ...)
 + 
 +  Optional input arguments should come in key-value pairs and can include
 +   '​noisecov' ​            = Nchan x Nchan matrix with noise covariance
 +   '​noiselambda' ​         = scalar value, regularisation parameter for the noise covariance matrix (default=0)
 +   '​filter_order' ​        = scalar, order of the mesh Butterwirth filter
 +   '​filter_bs' ​           = scalar, stop-band of the mesh Butterworth filter
 +   '​number_harmonics' ​    = Integer, number of mesh harmonics used (can be empty, the default will then be identity)
 +   '​lambda' ​              = scalar, regularisation parameter (can be empty, it will then be estimated from snr)
 +   '​snr' ​                 = scalar, signal to noise ratio
 +   '​reducerank' ​          = reduce the leadfield rank, can be '​no'​ or a number (e.g. 2)
 +   '​normalize' ​           = normalize the leadfield
 +   '​normalizeparam' ​      = parameter for depth normalization (default = 0.5)
 +   '​keepfilter' ​          = '​no'​ or '​yes',​ keep the spatial filter in the output
 +   '​prewhiten' ​           = '​no'​ or '​yes',​ prewhiten the leadfield matrix with the noise covariance matrix C
 +   '​scalesourcecov' ​      = '​no'​ or '​yes',​ scale the source covariance matrix R such that trace(leadfield*R*leadfield'​)/​trace(C)=1
 +   '​connected_components'​ = number of connected components of the source mesh (1 or 2)
 + 
 +  Note that leadfield normalization (depth regularisation) should be done by scaling
 +  the leadfields outside this function, e.g. in prepare_leadfield. Note also that
 +  with precomputed leadfields the normalization parameters will not have an effect.
 + 
 +  This implements
 +    Petrov Y (2012) Harmony: EEG/MEG Linear Inverse Source Reconstruction in the
 +    Anatomical Basis of Spherical Harmonics. PLoS ONE 7(10): e44439.
 +    doi:​10.1371/​journal.pone.0044439
 +</​pre></​html>​