Landau Lab
Draft for review BeaQuant real-time autoradiography · explainer and page · September 2026 · not yet final

Methods/Real-time autoradiography

Counting every particle, as it happens

The BeaQuant detector counts the particles given off by a radiolabelled brain section, one by one, and builds a quantitative image of where a tracer binds while we watch. It complements PET: PET follows a target in the living brain, and autoradiography maps it in brain tissue, in far finer detail.

0:48 · music, no narration · captions on screen, works with the sound off Detector and counting: simulatedAutoradiograms: published data

1 by 1

each β particle is counted individually, with its position

~2 hours

to read a [3H]UCB-J section in our studies. The same tracer took 5 weeks on film.

~20 µm

stated resolution for tritium, versus millimetres for PET

5 decades

linear range: twice the bound tracer gives twice the counts

01 How it works

From a slice of brain to a number per square millimetre

Autoradiography maps where a radioactive tracer has bound in a thin slice of tissue. Older methods record that radioactivity on film or on a phosphor plate over days or weeks. The BeaQuant instead detects the particles themselves, one at a time, in a gas-filled detector.

Photo to come

The BeaQuant at the PET Preclinical Facility, Aarhus University Hospital, with slides loaded

  1. Step 1A thin section and a tritium tracer

    The brain is cut into sections 20 micrometres thick, far thinner than a hair, and each section is mounted on a glass slide. The slide is incubated with a tracer that carries tritium (3H), a weakly radioactive form of hydrogen. The tracer binds its target. [3H]UCB-J, for example, binds SV2A, a protein on the vesicles that release neurotransmitters at synapses. Unbound tracer is washed away.

  2. Step 2Each decay sends out a β particle

    Every time a tritium atom decays, it gives off a β particle, which is a fast electron. Tritium's β particles are very weak: in tissue they travel no more than about 6 micrometres. The detector is built to catch them where they leave the tissue.

  3. Step 3Counted, one by one, in a gas

    The slide goes face down into the detector, over a thin layer of gas: neon with a little carbon dioxide. A β particle leaving the tissue knocks electrons out of the gas. A strong electric field multiplies them into an avalanche right at the tissue surface. That is why the position stays sharp. The electrons then drift through a fine metal mesh to a second amplification stage and a segmented anode, which records where the particle came from. Each particle becomes one count, with a position.

  4. Step 4The image builds up in real time

    Counts are shown as they arrive, so the image sharpens on screen during the scan. It can be stopped as soon as it is clear enough. In our studies, a [3H]UCB-J section is typically read in about 2 hours. Tracers that give a weaker signal are read for longer, up to about 22 hours.

  5. Step 5A number for every region

    The detector counts particles directly, so the signal rises in proportion to the amount of bound tracer, over about five orders of magnitude. Regions are read out in counts per second, or per minute, per square millimetre. Standards of known radioactivity are read alongside the sections to turn counts into concentrations. Binding in sections incubated with an excess of a competing drug is subtracted to give specific binding.

02 Alongside PET

Complementary to PET: same target, different measurement

Real-time autoradiography complements PET rather than replacing it. Both use radiolabelled ligands, often the same molecule with a different label: [11C]UCB-J for PET, [3H]UCB-J for tissue. But they measure different things. The data, and what we can conclude from them, can differ even when the results look alike.

PET · the living brainAutoradiography · brain tissue
MeasuresHow a tracer is taken up and bound in the brain over time. Kinetic modelling turns this into binding measures such as VT.Tracer bound to thin sections after incubation and washing, in counts or concentration per mm².
ConditionsThe tracer has to reach the brain. Blood flow, metabolism and the brain's own neurotransmitters can all affect the signal.Set by us: tracer concentration, buffer and time. A pre-wash removes most of the brain's own neurotransmitters.
TimeRepeated in the same animal, before and after a treatment or as a disease develops.One time point per brain, but many regions, tracers and animals in the same run.
DetailMillimetres. Signals average over whole regions.Tens of micrometres. Small nuclei and layers can be told apart.
Best forFollowing change in the living brain, and whether a drug reaches its target. The same method is used in people.Where exactly a target is, whether binding is specific, and how it relates to the anatomy of the same section.

Sometimes the two agree closely. In rats with a striatal lesion, [11C]UCB-J PET and [3H]UCB-J autoradiography found closely matching losses of SV2A binding (Thomsen et al. 2021). Sometimes the agreement is more limited. In minipigs, the two measures correlated across regions (r² = 0.72), but only because of the contrast with white matter; without it, the correlation disappeared (r² = 0.03; Thomsen et al. 2020). This is why we check one method against the other, region by region (see Methods).

Practical advantages in tissue

  • Many sections per runUp to 18 standard slides at once (manufacturer's figure). Each slide carries several sections.
  • Fine detailTens of micrometres instead of millimetres, so small nuclei and cortical layers can be measured.
  • Real-time read-outThe image is visible as it builds. A [3H]UCB-J read takes about 2 hours, not weeks.
  • Specificity in the same runTotal binding and non-specific binding (with a blocking drug) are read on adjacent sections.
  • Many targets, same brainsAdjacent sections can carry different tracers. One exercise study in rats compared four.
  • Tracers that aren't PET tracersTritium-labelled ligands work only in tissue. They exist for targets without a PET tracer, and can be custom-labelled.
  • No cyclotron neededMany tritium ligands can be bought off the shelf, and tritium's 12-year half-life removes the race against decay.
  • Histology on the same sectionAfter reading, sections can be Nissl-stained or used for protein analysis.

Collaborations

Many of our collaborations run through it

The BeaQuant is at the PET Preclinical Facility of the Department of Nuclear Medicine & PET Centre, Aarhus University Hospital.

See the tracer toolbox
Simulated counts · based on a published image

03 Tracer toolbox

Tritium tracers we work with

These tritium-labelled tracers are used for autoradiography only; tritium can't be imaged with PET. Each one maps a different part of brain biology in tissue sections.

  • [3H]UCB-JSV2AA protein on synaptic vesicles, used as a measure of synaptic density.
  • [3H]PK11195TSPOThe translocator protein. It rises when glial cells are activated, so it is used as a marker of neuroinflammation.
  • [3H]BU99008Imidazoline-2 binding sitesAn astrocyte-associated marker.
  • [3H]BCPP-EFMitochondrial complex IThe first enzyme of the mitochondrial respiratory chain, a window on energy metabolism.
  • [3H]DAMGOμ-opioid receptorThe receptor for the body's own opioids and for drugs such as morphine.
  • [3H]MPEPmGluR5Metabotropic glutamate receptor 5, part of glutamate signalling.
  • [3H]racloprideDopamine D2/3 receptorsKey receptors for dopamine, central to movement, reward and motivation.
  • [3H]CVN424 customGPR6An orphan receptor of the striatum, labelled specially for our GPR6 work.

Our published BeaQuant work has also used [3H]SR141716A (CB1), [3H]GBR-12935 (dopamine transporter), [3H]muscimol (GABAA), [3H]U69,593 and [3H]DPDPE (κ and δ opioid) and [3H]RS-127445 (5-HT2B). See the papers below.

Collaborate

Is there a question one of these tracers could answer, or a target you'd like to map in tissue?

Get in touch

04 Compared with film and plates

Hours instead of weeks, and a live image to decide when to stop

Film records radioactivity chemically, and you only see the result after it is developed. Phosphor plates store it, and are scanned afterwards. In the lab's earlier papers, tritium exposures took 4 to 5 weeks on film and 7 days on phosphor plates. In a 2019 study of human brain tissue, [3H]UCB-J needed 5 weeks on film. On the BeaQuant, our [3H]UCB-J sections are read in about 2 hours.

Nothing is exposed or developed, so the section itself is left intact. In several of our studies, the same sections were afterwards stained to outline anatomy, or used for protein analysis.

FilmPhosphor plateBeaQuant (gas detector)
RecordsDarkening of an emulsion, developed at the endStored energy, read by a laser scanner at the endEach particle, counted as it arrives
See the imageAfter developingAfter scanningDuring the scan, in real time
Tritium in our papers4–5 weeks7 days2 h for [3H]UCB-J · up to 22 h for other tracers
Linear rangeAbout 2 orders of magnitude4–5 ordersAbout 5 orders
AlsoCan saturate, so bright regions lose detailReusable plates; the signal fades during very long exposuresNot affected by γ-rays; can tell α from β particles

Sources: Delayre et al., Sci Rep 2020 (Table 1 and text); Donnard et al., JINST 2009; exposure times from the lab's papers listed below. On a prototype of this detector, the same sample needed 8 weeks on film and 15 hours on the detector (Donnard et al. 2009). Resolution depends on the isotope and the sample: for carbon-14 in rock, phosphor plates showed finer detail (Delayre et al. 2020).

05 In our research

What the images look like

Bar chart of specific [3H]UCB-J binding in counts per second per square millimetre in six minipig brain regions for control and sucrose groups, with prefrontal cortex lower after sucrose (p = 0.011); below, two rainbow-coloured BeaQuant autoradiograms of folded prefrontal cortex labelled control and sucrose.
The image at the end of the video. [3H]UCB-J binding to SV2A in Göttingen minipigs that drank 2 litres of 25% sucrose solution a day for 12 days, compared with controls. The results are in counts per second per mm² (top). In the prefrontal cortex, binding was lower after sucrose (p = 0.011, one-tailed). The autoradiograms (bottom) have no calibrated colour scale. Bærentzen SL, Thomsen MB, Alstrup AKO, Wegener G, Brooks DJ, Winterdahl M, Landau AM. Neuropharmacology 2024;256:110018, Fig. 2 · doi:10.1016/j.neuropharm.2024.110018 · CC BY 4.0, resized.
Three rows of rat brain autoradiograms, twelve coronal sections each, for three opioid receptor tracers, [3H]U69,593, [3H]DAMGO and [3H]DPDPE, with a column of non-specific binding sections beside each and colour bars in counts per minute per square millimetre from 0 to 20.
Three opioid receptors, one method. BeaQuant autoradiograms of κ ([3H]U69,593), μ ([3H]DAMGO) and δ ([3H]DPDPE) opioid receptor binding across the rat brain, from front to back. The NS columns show non-specific binding. The colour bars are in cp/min/mm². Krogsbaek M et al. J Chem Neuroanat 2023;132:102324, Fig. 3 · doi:10.1016/j.jchemneu.2023.102324 · CC BY 4.0, resized.
Top: twelve rat brain autoradiograms of [3H]U69,593 binding with a colour bar from 2 to 20 counts per minute per square millimetre, beside four non-specific sections. Bottom: the same sections after Nissl staining, in pale blue, with a 5 millimetre scale bar.
Read first, then stain the same sections. κ-opioid receptor binding on the BeaQuant (top). The same sections were then Nissl-stained (bottom) to outline small hypothalamic nuclei precisely. Scale bar: 5 mm. Krogsbaek M et al. J Chem Neuroanat 2023;127:102205, Fig. 5A–B · doi:10.1016/j.jchemneu.2022.102205 · CC BY 4.0, cropped to panels A–B.

The BeaQuant runs through much of our published work, in rats, mice, minipigs and human brain tissue. It has been used in studies of:

Synaptic density (SV2A)Parkinson's disease & exerciseDepression models & antidepressantsCannabinoid & glutamate receptorsOpioid receptors & feedingNeuroinflammation (TSPO)Diet & the minipig brainLiver disease & cognitionAlzheimer's disease22q11 deletionCocaine in adolescence

06 Papers using it

Our published work with the BeaQuant

Every paper below names the BeaQuant in its Methods. All tracers are tritium-labelled.

  1. 2025
    Agomelatine normalizes region-specific, diurnal mGluR5 dysregulation in a chronic mild stress rat model of depressionMPEPy · UCB-J
    Knudsen C, Thomsen MB, Højgaard K, … Elfving B · Prog Neuropsychopharmacol Biol Psychiatry
  2. 2025
    Chronic olanzapine treatment leads to increased opioid receptor expression and changes in feeding regulating neurons in the female rat hypothalamusDAMGO · U69,593 · DPDPE
    Krogsbaek M, Larsen NY, Yarmahmoudi F, … Nyengaard JR · Psychoneuroendocrinology
  3. 2025
    Modulation of the endocannabinoid system by (S)-ketamine in an animal model of depressionSR141716A
    Silva NR, Arjmand S, Domingos LB, … Joca SRL · Pharmacological Research
  4. 2025
  5. 2024
    Excessive sucrose consumption reduces synaptic density and increases cannabinoid receptors in Göttingen minipigsUCB-J · raclopride · MPEPγ · SR141716A
    Bærentzen SL, Thomsen MB, Alstrup AKO, … Landau AM · Neuropharmacology
  6. 2024
  7. 2024
  8. 2024
    A single dose of cocaine raises SV2A density in hippocampus of adolescent ratsUCB-J · GBR-12935
    Rossi R, Bærentzen SL, Thomsen MB, … Landau AM · Acta Neuropsychiatrica
  9. 2024
  10. 2023
  11. 2023
    Changes in hypothalamic mu-opioid receptor expression following acute olanzapine treatment in female rats: implications for feeding behaviorDAMGO · U69,593 · DPDPE
    Krogsbaek M, Larsen NY, Landau AM, … Nyengaard JR · Journal of Chemical Neuroanatomy
  12. 2023
    Spatial quantification of single cell mRNA and ligand binding of the kappa opioid receptor in the rat hypothalamusU69,593
    Krogsbaek M, Larsen NY, Landau AM, … Nyengaard JR · Journal of Chemical Neuroanatomy
  13. 2022
  14. 2021
    Exercise protects synaptic density in a rat model of Parkinson's diseaseUCB-J · raclopride · PK11195 · DAMGO
    Binda KH, Lillethorup TP, Real CC, … Landau AM · Experimental Neurology
  15. 2021
    In vivo imaging of synaptic SV2A protein density in healthy and striatal-lesioned rats with [11C]UCB-J PETUCB-J
    Thomsen MB, Jacobsen J, Lillethorup TP, … Landau AM · Journal of Cerebral Blood Flow & Metabolism
  16. 2021
    α-Synuclein overexpression increases dopamine D2/3 receptor binding and immune activation in a model of early Parkinson's diseaseUCB-J; other tracers on phosphor plates
    Stokholm K, Thomsen MB, Phan JA, … Landau AM · Biomedicines
  17. 2020
    Preclinical PET studies of [11C]UCB-J binding in minipig brainUCB-J
    Thomsen MB, Schacht AC, Alstrup AKO, … Landau AM · Molecular Imaging and Biology

All publications: publication list →

07 Technical notes

For the specialists

Detector, performance and sources
Instrument
BeaQuant real-time digital autoradiography system (ai4R, Nantes, France). Acquisition runs in Beavacq and analysis in Beamage.
Detector
A micro-pattern gaseous detector with a parallel ionization multiplier (PIM), based on Micromegas.
The slide, with conductive tape on its back, is the cathode and faces down. A first amplification gap of about 200 µm sits directly against the section; this reduces parallax.
Next come a drift gap of about 1 cm, a second micromesh with a 50 µm amplification gap, and a segmented anode read out on two axes. The detector counts pulses.
Gas
Neon with 10% CO2. It is non-toxic and non-flammable.
Detects
β and α particles directly. It also detects Auger electrons, for example from 125I. It is insensitive to γ-rays, and energy information allows α and β particles to be told apart. The manufacturer states that it works with all radioisotopes.
Resolution
Manufacturer's figures: 20 µm for β (tritium), 10 µm for α, and 50 µm for high-energy β and β+. The α figure differs between ai4R pages.
Published measurements: about 20 µm for 3H, and 30 µm FWHM for 3H on the 2009 prototype.
Our images are usually analysed at 100 µm pixels.
Range
Linear from 5 × 10−4 to about 900 cpm/mm². The maximum is 30,000 counts per second over the whole field.
Stated sensitivity: 0.0005 cpm/mm².
In our hands, tritium standards correlate with BeaQuant counts at R² = 0.999.
Isotopes
Mostly tritium: see the tracer toolbox. We have also imaged fluorine-18 on the BeaQuant. For shorter-lived isotopes we more often use a Typhoon phosphor-plate reader at the PET Centre; our [18F]PSMA-1007 study is an example.
Practical
The sample sits at high voltage. Plastic tritium standards can discharge and create artefacts, so we calibrate with standard slides made from blood spiked with known tritium concentrations.
  • Donnard J, Arlicot N, Berny R, et al. Advancements of labelled radio-pharmaceutics imaging with the PIM-MPGD. JINST 2009;4:P11022 · doi:10.1088/1748-0221/4/11/P11022
  • Donnard J, Thers D, Servagent N, Luquin L. High spatial resolution in β-imaging with a PIM device. IEEE Trans Nucl Sci 2009;56:197–200 · IEEE Xplore
  • Delayre C, Sammaljärvi J, Billon S, et al. Comparison of phosphor screen autoradiography and micro-pattern gas detector based autoradiography for the porosity of altered rocks. Sci Rep 2020;10:9455 · doi:10.1038/s41598-020-65791-7
  • Ang JWL, Bongrand A, Duval S, et al. Detecting radioactive particles in complex environmental samples using real-time autoradiography. Sci Rep 2024 · doi:10.1038/s41598-024-52876-w
  • ai4R. BeaQuant product information · ai4r.com/beaquant · TNU, Aarhus University: Beaquant real-time autoradiography

Questions for review

  1. Tracer toolbox: are the one-line descriptions of each target right? In particular, is "astrocyte-associated marker" the wording you want for [3H]BU99008?
  2. Photo: the placeholder under "How it works" is ready for your new photo of the instrument.
  3. Collaborations: should we name collaborating groups, or keep to research topics as now? The earlier wording "we are its main users" has been left out; say if you'd like it back.
  4. Comparison table: does the PET-versus-autoradiography table say what you want about what each method measures?
  5. Video: the ending "Two views of one target" pairs [11C]UCB-J PET with [3H]UCB-J autoradiography. A separate PET-versus-autoradiography comparison film is being planned.
MethodsFrom signal to evidenceHow PET, autoradiography, tissue analyses and behaviour check one another.All methods → FilmThe lab in 46 secondsFrom a single PET signal to the lab's data, methods and themes.Watch →