T1 · longitudinal recovery
Recovered along B0. At one T1: 63%.
Mz / M0 = 1 − e−t/T1
An interactive MRI learning lab
MRI, made visible.
Explore what the signal does, how scan timing changes an image, and why blood oxygenation can reveal changes associated with brain activity.
Recovery and decay happen together.
Start just after a 90° RF pulse. T1 describes recovery along the main field. T2 describes loss of transverse coherence. T2* includes additional dephasing from magnetic-field differences.
Change a time constant to change the rate.
T2 · transverse decay
Transverse signal remaining. At one T2: 37%.
Mxy / Mxy(0) = e−t/T2
T2* · extra dephasing
1/T2* = 1/T2 + 1/T2′
Phase arrows are schematic; percentages and curves follow the exponential model. Graphs use different time ranges. Dashed vertical lines mark your selected time when it is in range. In the last graph, dashed orange is T2 and solid purple is T2*.
Why the star matters: an ideal 180° RF pulse refocuses dephasing from static field offsets to form a spin echo governed by T2. It does not recover intrinsic T2 decay. Gradient echoes retain T2* sensitivity; conventional BOLD fMRI commonly uses them.
Try a comparison
Shorten T2′: T2* falls faster. Lengthen T2′: T2* approaches T2. T1 recovery happens simultaneously. These controls are independent mathematical parameters; not every combination describes a real tissue.
The same tissue. Different timing.
TR is the interval between excitation pulses. TE is the time from excitation to the measured echo. Keep tissue constants fixed and change the scanner's timing.
Timing schematic; spacing is not to scale. Gradient echo uses gradient reversal, without a 180° RF refocusing pulse.
Illustrative tissue map
A schematic, not a patient scan. A fixed brightness scale makes overall signal loss visible.
What creates each signal?
Model and illustrative tissue values
The recovery term sets the magnetization available for excitation. The decay term sets how much transverse signal survives to TE. Longer TR reduces T1 saturation; longer TE increases transverse-decay sensitivity while reducing signal.
Illustrative constants in ms: white matter T1 800, T2 90, T2* 45; gray matter 1300, 100, 50; fluid 4000, 2000, 1000. Equal proton density isolates relaxation. Real values depend on tissue and field strength.
The model assumes repeated 90° excitation with ideal spoiled recovery. The gradient-echo option substitutes T2* to isolate its effect; real gradient-echo signals also depend on flip angle and sequence details. Presets illustrate contrast, not scanner protocols.
From neural activity to a BOLD signal.
BOLD means blood-oxygen-level dependent. It measures a vascular response associated with neural activity, rather than recording neurons firing directly.
1 · ActivityA brief event increases local energy demand.
2 · Blood flowFlow typically increases more than oxygen consumption.
3 · DeoxyhemoglobinLocal deoxygenated hemoglobin decreases during the positive response.
4 · T2*Less susceptibility-driven dephasing means longer effective T2*.
5 · BOLDMore signal survives to TE. Repeated volumes capture a delayed change.
The response arrives later
Continuous model● Acquired volumesDashed orange: event
An illustrative single-event response, delayed peak and small undershoot. Shape and delay vary across people and regions.
OxygenatedDeoxygenated
Cell colors illustrate the mechanism; counts are not physiological measurements.
Longer TE increases modeled BOLD percentage sensitivity, but baseline signal decays more. Shorter TR adds samples; it does not remove the vascular delay.
How this becomes an activation map
The scanner repeats a T2*-sensitive acquisition, often gradient-echo echo-planar imaging, to collect a time series at each voxel. Researchers model task timing with a hemodynamic response, account for motion and other confounds, and estimate task-related responses. Statistical maps are commonly overlaid on an anatomical scan.
This model fixes baseline T2* at 40 ms and illustrates a change in R2* = 1/T2*. TE affects exp(−TE × R2*). Here TR controls sampling only; saturation, noise and physiological variability are omitted. BOLD reflects neural–vascular coupling, not a unique measure of firing rate.
How signals acquire an address.
Gradients make frequency and phase depend on position. Columns have different frequencies; rows have different starting phases. The receiver measures their sum. Many encoding measurements and an inverse Fourier transform reconstruct the image.
Each k-space measurement mixes contributions across many positions. One k-space point is not one voxel.