⚠ Note on SWI diagrams: SWI is a T2*-weighted 3D spoiled GRE sequence. There is no T1-weighted or conventional T2-weighted mode — the weighting is always T2*. The four diagrams show T2* contrast at the two clinically relevant TE values at 1.5T and 3T. The key physics difference from SE/TSE/IR: Mxy decays as T2* (not T2) — faster, because field inhomogeneities are not refocused. The T2* values shown include both intrinsic T2 and susceptibility-driven dephasing.
SWI — T2*-weighted GRE · 1.5T · Optimal TE
TR = 40 ms · TE = 40 ms · α = 18° · 5 cycles — TE≈T2*_GM maximises susceptibility contrast; 4 TR cycles shown for steady-state
Fat (T1=260ms, T2*=40ms at 1.5T)
White matter (T1=650ms, T2*=70ms)
Grey matter (T1=950ms, T2*=60ms)
Muscle (T1=860ms, T2*=25ms)
CSF (T1=3600ms, T2*=1500ms)
Venous blood (T2*=5–15ms — very dark)
Caption — SWI T2*-weighted (1.5T, TE=40ms): SWI uses a small flip angle (α=18°) spoiled GRE. After each α pulse, Mz recovers via T1 relaxation toward the steady-state value (Ernst angle regime). At TR=40ms and α=18°, the steady-state Mz is a fraction of M₀ — the sequence runs in a quasi-steady state. The key difference from SE/TSE is Mxy decay: without a 180° refocusing pulse, transverse magnetisation decays at the T2* rate (T2* = T2 combined with field inhomogeneity). At TE=40ms: muscle (T2*=25ms) → e^(−40/25)≈20% signal; grey matter (T2*=60ms) → e^(−40/60)≈51% signal; white matter (T2*=70ms) → e^(−40/70)≈57% signal; CSF (T2*=1500ms) → e^(−40/1500)≈97% signal. Venous blood containing deoxyhaemoglobin (T2*≈5–15ms) decays to near-zero — this is the basis for venous contrast in SWI. Any iron-containing structure (haemosiderin, T2*<10ms) also appears near-dark. Calcifications produce susceptibility dephasing and appear dark on magnitude (but bright on phase).
SWI — T2*-weighted GRE · 1.5T · Extended TE
TR = 50 ms · TE = 55 ms · α = 18° · 4 cycles — longer TE → more T2* contrast; lower SNR; microbleed blooming increases
Caption — SWI extended TE (1.5T, TE=55ms): At TE=55ms, T2* contrast is further amplified. Muscle (T2*=25ms) → e^(−55/25)≈11% — near-zero; grey matter (T2*=60ms) → e^(−55/60)≈40%; white matter (T2*=70ms) → e^(−55/70)≈46%. CSF remains bright. At this TE, SNR is lower than at TE=40ms. The susceptibility effect of microbleeds and haemosiderin is maximised, but the blooming artefact radius also increases proportionally with TE. Clinical practice uses TE=35–45ms at 1.5T as the optimal balance.
SWI — T2*-weighted GRE · 3T · Optimal TE
TR = 28 ms · TE = 25 ms · α = 18° · 5 cycles — at 3T, T2* values are shorter; TE≈20–25ms provides equivalent susceptibility contrast to TE=40ms at 1.5T
Fat (T1=380ms, T2*=28ms at 3T)
White matter (T1=830ms, T2*=55ms)
Grey matter (T1=1300ms, T2*=45ms)
Muscle (T1=1412ms, T2*=20ms)
CSF (T1=4000ms, T2*=1200ms)
Venous blood (T2*=3–8ms at 3T — even darker)
Caption — SWI at 3T (TE=25ms): At 3T, T2* values are shorter than at 1.5T because the susceptibility effects scale with B0. Grey matter T2*≈45ms (vs 60ms at 1.5T); WM T2*≈55ms (vs 70ms at 1.5T); venous blood T2*≈3–8ms (vs 5–15ms at 1.5T — even more extreme at 3T). At TE=25ms: grey matter signal ≈ e^(−25/45)≈57%; white matter ≈ e^(−25/55)≈63%; muscle (T2*=20ms) ≈ e^(−25/20)≈29% — already dark; venous blood ≈ near-zero. The 3T SWI has twice the susceptibility sensitivity of 1.5T for the same TE — this is why microbleed detection is substantially more sensitive at 3T (smaller/weaker susceptibility sources are detectable). The shorter T2* also means phase accumulates twice as fast, allowing shorter TE for equivalent phase contrast.
SWI — T2*-weighted GRE · 3T · Extended TE
TR = 35 ms · TE = 35 ms · α = 18° · 4 cycles — at 3T, TE=35ms approaches T2* of GM; high susceptibility contrast but significant blooming risk near implants
Caption — SWI extended TE at 3T: At TE=35ms at 3T, both muscle and fat have decayed substantially. Grey matter ≈ e^(−35/45)≈46%; white matter ≈ e^(−35/55)≈53%. The susceptibility effect of microbleeds at 3T with TE=35ms is maximal but the blooming radius is also largest. Near metallic implants, TE=35ms at 3T produces artefact that can extend 2–4 cm from the implant surface, rendering SWI non-interpretable in that region. Clinical SWI at 3T typically uses TE=20–30ms as the optimal balance between susceptibility sensitivity and artefact control.