fmri-task_spm_firstlevel_auditory

Task fMRI: SPM25 + FSL FEAT Auditory Replication

Goal: Reproduce the classic SPM Auditory first-level GLM in two software ecosystems and compare the same Listening > Rest contrast.

Snapshot

• Dataset: SPM Auditory / MoAEpilot
• Subject: sub-01
• Volumes: 84
• TR: 7 s
• Tools: MATLAB R2025b, SPM25, FSL FEAT 6.0.7.12
• Status: complete
• Last updated: 2026-07-15

SPM25 analysis

• Realignment and reslicing
• Descending slice timing, 64 slices, reference slice 32
• T1-to-EPI coregistration
• Segmentation and MNI normalization
• 3 mm isotropic functional output
• 6 mm FWHM smoothing
• First-level GLM with six motion nuisance regressors
• Contrast: Listening > Rest
• Inference: voxel-level whole-brain FWE p < 0.05
• Observed threshold: T > 5.296299

FSL FEAT replication

• MCFLIRT motion correction
• Regular-descending slice timing
• BET brain extraction
• SUSAN smoothing, 6 mm FWHM
• High-pass filter, 128 s
• FILM prewhitening
• Double-gamma HRF
• Six motion nuisance EVs
• FLIRT registration to T1 and MNI152
• Contrast: Listening > Rest
• Inference: cluster-forming Z > 3.1, cluster significance p < 0.05

Results

SPM motion QC

SPM thresholded activation

SPM maximum-intensity projections

FSL FEAT thresholded activation

Cross-software summary

The SPM and FEAT results are compared qualitatively at the pattern level. Their T/Z statistics, cluster sizes, and significance values are not directly interchangeable because the inferential procedures differ.

Peak-coordinate correspondence is approximate because SPM MNI and FSL MNI152 normalized coordinates are not guaranteed to represent identical template geometry; no voxelwise overlap analysis is claimed.

Statistical summary

SPM25

• Largest clusters: 403 and 213 voxels
• Peaks: [-63, -28, 14] and [57, -22, 11]
• Peak T values: approximately 11.87

FSL FEAT

• Left dominant cluster: 863 voxels, Z max 10.5, peak [-61.1, -27.0, 8.85]
• Right dominant cluster: 718 voxels, Z max 10.3, peak [58.7, -22.8, 5.68]

Both pipelines recovered strong bilateral auditory activation. This is a conceptual replication rather than an expectation of exact voxelwise equivalence because the packages use different temporal models, registration strategies, statistical images, and inference procedures.

Reproducibility and archive status

The repository preserves completed-result figures, quantitative tables, threshold information, software metadata, and data provenance. The original executable analysis scripts and complete SPM/FEAT derivative workspaces are not present in the audited archive; this is therefore not presented as a fully executable end-to-end pipeline.

Archived resources:

• results/tables/table1_spm_peaks.csv
• results/tables/table2_fsl_clusters.txt
• results/tables/spm_threshold_info.txt
• results/tables/fsl_threshold_info.txt
• env/TOOL_VERSIONS.md
• DATA_SOURCES.md

Limitations

Single-subject tutorial analysis, no group-level inference, affine FSL standard-space registration, and non-equivalent cross-package thresholding procedures.

---

Author: Rene Andrade Rey · 🧪 ORCID: https://orcid.org/0000-0001-5627-579X · 🌐 Scholar: https://scholar.google.es/citations?hl=es&user=Nl3ApFEAAAAJ

This site is open source. Improve this page.