Compute Pairwise Riemannian Distances Between Subject SPD Representations
Source:R/riemannian_methods_hatsa.R
riemannian_distance_matrix_spd.RdCalculates a matrix of Riemannian distances (Log-Euclidean metric) between subjects, based on specified SPD matrix representations derived from the HATSA object (e.g., covariance of aligned coefficients or various FC matrices).
Usage
riemannian_distance_matrix_spd(object, ...)
# S3 method for class 'hatsa_projector'
riemannian_distance_matrix_spd(
object,
type = "cov_coeffs",
subject_data_list = NULL,
spd_regularize_epsilon = 1e-06,
k_conn_params = NULL,
...
)Arguments
- object
A `hatsa_projector` or `task_hatsa_projector` object, or directly a list of SPD matrices.
- ...
Additional arguments passed to `riemannian_distance_spd` or to `get_spd_representations` if used.
- type
Character string indicating the type of SPD representation to use. Commonly `"cov_coeffs"` (default) for the `k x k` covariance of aligned spectral coefficients. Other types (e.g., `"fc_conn"`, `"fc_task"`) would rely on `get_spd_representations` (from RGEOM-003) to provide the SPD matrices.
- subject_data_list
Optional. A list of subject time-series matrices (T_i x V_p). May be needed by `get_spd_representations` for certain `type` values.
- spd_regularize_epsilon
Epsilon for regularizing SPD matrices.
- k_conn_params
Parameters for connectivity calculation if type involves FC, passed to `get_spd_representations`.