KAHIBARO
25.1. Time in GATE
Table of Contents
Simulation intervals
Time in GATE represents the evolution of your simulated acquisition, not the wall‑clock time on your computer. Every primary particle and every interaction is tagged with a simulation time value, measured in seconds, that you control through the simulation configuration.
At the most basic level you define a total time interval during which particles are allowed to be generated. Within this interval GATE assigns a specific time to each primary particle according to the source activity and time settings. When you configure a source with an activity, GATE automatically converts that to an average emission rate and samples primary times as a Poisson process over the defined interval. If you do not use activity, but instead specify only a fixed number of particles, you can still assign a time interval and GATE will distribute those particles over that interval in a simple way.
The simulation interval is typically defined by a start time, often taken as $t = 0$, and an end time $t = T_\text{end}$. All time dependent changes in the simulation are expressed relative to this axis. Moving geometries, changing activities, and other dynamic elements are defined as functions of time that are evaluated at the current simulation time of each event.
In practice, the simulation interval controls when time‑dependent components are updated. For example, a rotating detector can be configured so that its rotation angle depends on the current simulation time, such as $\theta(t) = \omega \, t$, where $\omega$ is an angular speed. A respiratory motion model for a patient phantom may specify organ positions as periodic functions of time, like $x(t) = x_0 + A \sin(2\pi t / T_\text{breath})$. Whenever an event is generated at time $t$, GATE evaluates these expressions and updates the position or orientation of the corresponding volumes before transporting the particles.
It is important to remember that the transport of a single event can involve internal time scales, for example the time of flight of particles and delayed processes. These internal times are handled by the physics engine, and the event maintains its own absolute time that is the sum of its start time and the intra‑event physical delays. Time‑of‑flight information in detectors, coincidence timing, and decay chains all use this consistent time reference.
In dynamic simulations you may need to define several distinct intervals to describe different acquisition phases, such as list‑mode frames or gates. Conceptually, each frame corresponds to a sub‑interval of the global simulation time where particular conditions apply, for instance a static detector position or a specific activity level. By choosing these intervals carefully you can mimic clinical protocols such as gated cardiac imaging or multi‑phase therapy delivery.
The simulation time in GATE is a physical time axis that controls source emission, motion, and dynamic behavior. All time‑dependent configurations, such as moving geometries or changing activities, must be expressed as explicit functions of this simulation time and must use consistent time units (for example s, ms, ns).
Acquisition duration
The acquisition duration is the total length of time over which your virtual scanner or treatment system is simulated. It is directly linked to the simulation interval but focuses on how long the system is considered to be "on" and collecting data.
For imaging simulations the acquisition duration is chosen to reflect realistic scan times. In PET, you might simulate several minutes per bed position; in SPECT or CT you might simulate shorter durations per projection angle. In therapy simulations the duration can represent beam‑on time, fraction delivery time, or the time window of a dynamic arc.
When you specify an acquisition duration together with a source activity, GATE uses both to determine the total expected number of decays. For a constant activity $A$ and duration $T$, the expected number of decays is $N \approx A \times T$. Monte Carlo sampling introduces statistical fluctuations, so the actual number of emitted particles varies from run to run, but it remains centered around this expectation. If you want a fixed number of primaries instead, you typically configure the source with a particle count and choose an acquisition duration that corresponds to the experimental scenario, so that time stamps and count rates remain realistic.
Acquisition duration also controls the time window over which moving components evolve. For example, a SPECT detector rotating at a constant angular speed over 360 degrees during a 20 minute acquisition will have its instantaneous angle linked to the current simulation time, from $t = 0$ to $t = 1200$ s. Similarly, in dynamic imaging or 4D CT, each temporal frame corresponds to a shorter acquisition duration within the overall scan, and you can treat these as separate simulations or as segments of a single continuous simulation, depending on how you configure sources and actors.
The choice of acquisition duration has a direct impact on the statistical quality of your results. Longer durations mean more emitted particles and therefore lower statistical uncertainty in tallies such as dose or detector counts, at the cost of longer computation time. Shorter durations may match a clinical protocol but yield noisier results; in such cases you often compensate by running multiple independent simulations and combining the outputs.
In time‑resolved analyses, such as time‑activity curves or time‑of‑flight histograms, the acquisition duration defines the global limits of the histograms you can construct. Within this global window you can define finer temporal bins, for example 1 s, 100 ms, or even smaller, and use the event times assigned by GATE to fill these bins.
Always choose the acquisition duration to match the physical scenario you want to model. Combine it consistently with source activity and motion definitions, and ensure that all time‑dependent configurations and analysis bins lie within this duration and use compatible time units.
Views: 13