|
Smallest intent_code that indicates a statistic. |
|
[C2, chap 25] Beta distribution (2 params): p1=a, p2=b. Density(x) proportional to x^(a-1) * (1-x)^(b-1). |
|
[U, chap 3] Binomial distribution (2 params): p1 = number of trials, p2 = probability per trial. Prob(x) = (p1 choose x) * p2^x * (1-p2)^(p1-x), for x=0,1,...,p1. |
|
[C1, chap 18] Chi distribution (1 param): p1 = DOF. Density(x) proportional to x^(p1-1) * exp(-x^2/2) for x > 0. p1 = 1 = 'half normal' distribution p1 = 2 = Rayleigh distribution p1 = 3 = Maxwell-Boltzmann distribution. |
|
[C1, chap 18] Chi-squared (1 param): p1 = DOF. Density(x) proportional to exp(-x/2) * x^(p1/2-1). |
|
[C2, chap 29] Noncentral chi-squared statistic (2 params): p1 = DOF, p2 = noncentrality parameter. |
|
Dimensionless value - no params - although, as in _ESTIMATE the name of the parameter may be stored in intent_name. |
|
To signify that the vector value at each voxel is to be taken as a displacement field or vector:
|
|
To signify that the value at each voxel is an estimate of some parameter, set intent_code = NIFTI_INTENT_ESTIMATE. The name of the parameter may be stored in intent_name. |
|
[C2, chap 22] Extreme value type I (2 params): p1 = location, p2 = scale cdf(x) = exp(-exp(-(x-p1)/p2)). |
|
[C2, chap 27] Fisher F statistic (2 params): p1 = numerator DOF, p2 = denominator DOF. |
|
[C2, chap 30] Noncentral F statistic (3 params): p1 = numerator DOF, p2 = denominator DOF, p3 = numerator noncentrality parameter. |
|
[C1, chap 17] Gamma distribution (2 params): p1 = shape, p2 = scale. Density(x) proportional to x^(p1-1) * exp(-p2*x). |
|
To store an M x N matrix at each voxel:
|
|
[C1, chap 15] Inverse Gaussian (2 params): p1 = mu, p2 = lambda Density(x) proportional to exp(-p2*(x-p1)^2/(2*p1^2*x)) / x^3 for x > 0. |
|
To signify that the value at each voxel is an index into some set of labels, set intent_code = NIFTI_INTENT_LABEL. The filename with the labels may stored in aux_file. |
|
[C2, chap 24] Laplace distribution (2 params): p1 = location, p2 = scale. Density(x) proportional to exp(-abs(x-p1)/p2). |
|
Data is log10(p-value) (no params). To be safe, a program should compute p = pow(10.,-abs(this_value)). The nifti_stats.c library returns this_value as positive, so that this_value = -log10(p). |
|
[C2, chap 23] Logistic distribution (2 params): p1 = location, p2 = scale. Density(x) proportional to sech^2((x-p1)/(2*p2)). |
|
Data is ln(p-value) (no params). To be safe, a program should compute p = exp(-abs(this_value)). The nifti_stats.c library returns this_value as positive, so that this_value = -log(p). |
|
To signify that the value at each voxel is an index into the NeuroNames labels set, set intent_code = NIFTI_INTENT_NEURONAME. |
|
To signify that the value at each location is a node index, from a complete surface dataset. |
|
[C1, chap 13] Normal distribution (2 params): p1 = mean, p2 = standard deviation. |
|
To signify that the vector value at each voxel is really a spatial coordinate (e.g., the vertices or nodes of a surface mesh):
|
|
[U, chap 4] Poisson distribution (1 param): p1 = mean. Prob(x) = exp(-p1) * p1^x / x! , for x=0,1,2,.... |
|
Data is a 'p-value' (no params). |
|
To signify that the vector value at each voxel is a quaternion:
|
|
To signify that the vector value at each location is a 4 valued RGBA vector, of whatever type.
|
|
To signify that the vector value at each location is an RGB triplet, of whatever type.
|
|
To signify that the value at each location is a shape value, such as the curvature. |
|
To store an NxN symmetric matrix at each voxel:
|
|
To signify that the value at each location is from a time series. |
|
To signify that the vector value at each voxel is really a triple of indexes (e.g., forming a triangle) from a pointset dataset:
|
|
[C2, chap 28] Student t statistic (1 param): p1 = DOF. |
|
[C2, chap 31] Noncentral t statistic (2 params): p1 = DOF, p2 = noncentrality parameter. |
|
[C2, chap 26] Uniform distribution: p1 = lower end, p2 = upper end. |
|
[C1, chap 21] Weibull distribution (3 params): p1 = location, p2 = scale, p3 = power. Density(x) proportional to ((x-p1)/p2)^(p3-1) * exp(-((x-p1)/p2)^p3) for x > p1. |
|
[C1, chap 13] Standard normal (0 params): Density = N(0,1). |
|
Largest intent_code that indicates a statistic. |