KAHIBARO
16.2. Hits
Table of Contents
Raw interaction information
When a particle passes through a detector volume in a GATE simulation, Geant4 follows its motion step by step. At each interaction with matter, physical processes such as Compton scattering, photoelectric absorption, or ionization can deposit energy and change the particle trajectory. The most basic description of what happens in the detector is the sequence of these individual interactions.
At the Geant4 level, every step contains detailed information such as pre and post step positions, the process that limited the step, and the step length. This raw level is very fine grained and is rarely used directly for detector studies. Instead, GATE groups step-level information inside a sensitive volume into more convenient summaries. These summaries are called hits.
A hit corresponds to one or more particle steps that satisfy the conditions of the detector’s sensitive region and are collected by a specific hit actor or sensitive detector. For each hit, GATE records a compact set of quantities that describe the interaction from the point of view of the detector. Typical information includes the deposited energy $E_{dep}$ in the sensitive volume, the position of the interaction, for example the 3D coordinates of the energy deposition, the global time when the interaction occurred, and identifiers that link the hit to the event, track, particle type, and detector volume.
Although a hit is already a processed object compared with raw Geant4 steps, it is still considered raw interaction information within the context of detector modeling. Hits describe what happened physically in the scintillator crystal or sensor, before any modeling of optical photons, electronics, timing response, or energy resolution. They sit exactly at the interface between physics transport and detector response.
In a typical GATE workflow, you configure which volumes are sensitive and which attributes are stored in hit collections. For example, a PET crystal volume can be declared as sensitive so that every energy deposition in it becomes part of a hit collection. If a particle deposits energy multiple times in the same crystal during one event, depending on your configuration the steps can be merged into a single hit or stored as separate hits. This allows flexibility in how closely you want to follow the physical interactions.
Hits are also linked to the Monte Carlo event structure. Each hit contains an event ID so that you can trace all hits produced by the same primary particle. Additional indices, such as track ID or parent ID, allow you to distinguish hits from different secondaries like scattered photons or electrons. This is particularly important in medical imaging simulations where you often need to classify events later as primary, scattered, or background.
Because hits are recorded before digitization, they are crucial for detailed detector studies. You can use hit information to validate geometry, check where and how much energy is deposited, and study how particles interact inside complex detector structures. For example, by plotting hit positions you can verify that interactions happen inside the intended volumes and not in dead regions or shielding.
Hits, however, can become very numerous. High statistics simulations, large voxelized phantoms, or fine segmentation can produce millions of hits per run. For absolute beginners it is important to understand that recording many detailed hit attributes increases output size and analysis time. You will typically choose a balance between the level of detail you need and the amount of data you can handle.
A hit is the fundamental unit of raw detector interaction in GATE: it records deposited energy, position, time, and identifiers in a sensitive volume, before any electronic or reconstruction effects are applied.
Detector hits
Detector hits connect the underlying particle transport with detector-specific structures such as crystals, modules, pixels, or strips. Each hit belongs to a particular detector element, identified by one or more volume indices or IDs. This mapping allows you to go from the physical coordinates of an interaction to the logical detector channel that would respond in a real system.
In practice, you assign sensitive properties to detector volumes so that GATE can automatically produce hit collections. For a PET scanner, you typically set the scintillation crystals as sensitive volumes. For a gamma camera, the NaI crystal is often the sensitive detector. In each case, GATE uses the volume hierarchy and copy numbers to determine which detector element receives the hit.
The most important quantity in a detector hit is the deposited energy $E_{dep}$. For ideal interactions, such as full absorption of a 511 keV photon in a PET crystal, the hit will record an energy close to the photon energy. For partial interactions such as Compton scatter, the hit will contain only a fraction of the photon energy. Because multiple interactions can occur in the same detector element during one event, hits may need to be combined at a later stage. That combination is handled by the digitizer, not at the hit level, so at the hits stage you usually want to keep the information for each interaction or crystal-volume accumulation.
Other common attributes in detector hits include the 3D position of the interaction, often stored as $x, y, z$ in the global coordinate system, the global time of the interaction, and IDs describing the detector hierarchy, such as crystal ID, module ID, or ring ID. Additional fields might contain the particle type code, event ID, track ID, and sometimes the name of the interaction process that caused the energy deposition.
It is useful to think of detector hits as a detailed map of where and when the detector material absorbed energy. This map is still idealized because it assumes perfect knowledge of the interaction physics, exact positions, and exact times. No detector imperfections, light transport, or electronics noise are included yet. As you move further along the simulation chain, later components like digitizers will apply energy blurring, spatial blurring, timing smearing, and thresholds to transform these idealized hits into realistic detector signals such as singles and coincidences.
For many analysis tasks, detector hits are the right level of abstraction to understand and debug the system. For example, you can study energy deposition patterns inside a PET ring, check how often photons interact in the front versus the back of a crystal, or see how shielding affects background interactions. Because each hit is tied to a volume and has a well-defined position, you can create occupancy maps, energy maps, or depth-of-interaction distributions directly from hit data.
In medical imaging and dosimetry contexts, detector hits are also a starting point for consistency checks with physics expectations. You can compare the sum of deposited energies in a given region with analytical attenuation estimates or known interaction probabilities. If the distributions of detector hits are not reasonable, there is usually a problem with geometry, materials, or physics settings that should be fixed before you trust later stages like digitized signals or images.
Finally, detector hits play an important role in separating physical effects. Since hits preserve information about the particle origin and interaction history, you can distinguish between primary photons and scattered photons, or identify hits due to background particles. This is essential for later classification of events into true, scattered, or random categories, especially in PET and SPECT simulations.
Detector hits represent ideal energy depositions in specific detector elements, tagged with position, time, and detector IDs. They form the bridge between Geant4 particle transport and the higher level detector response modeled by the digitizer.
Views: 9