Decay Time Constants (Tau) in Calcium Imaging Spike Inference Pipelines

Decay Time Constants (Tau) in Calcium Imaging Spike Inference Pipelines#

Most calcium imaging pipelines model neural activity as an exponential decay process following each spike, making τ a key hyperparameter.

This guide summarizes in vivo τ values used or recommended by leading spike inference algorithms — FOOPSI, OASIS, Suite2p, CaImAn, and CASCADE — as reported in peer-reviewed literature and official documentation:contentReference[oaicite:0]{index=0}.

FOOPSI (Fast Non-Negative Deconvolution)#

FOOPSI (Vogelstein 2010) introduced probabilistic inference of spike trains using an exponential calcium model:

\[C_t = \gamma C_{t-1} + A s_t\]

where ( \gamma = e^{-\Delta t / \tau} ) defines the decay rate.

In in vivo work, FOOPSI typically assumes τ ≈ 1.0 s for GCaMP3–GCaMP5 indicators:contentReference[oaicite:1]{index=1}, which display relatively slow kinetics.
This constant was fixed across datasets and used to calibrate spike likelihood.

OASIS and CNMF (Pnevmatikakis & Friedrich, 2016–2017)#

OASIS improved upon FOOPSI by introducing a convex optimization solver for AR(1)/AR(2) calcium models (Friedrich 2017).
τ is explicitly passed as a parameter in seconds, often derived from indicator kinetics:contentReference[oaicite:2]{index=2}.

Indicator	τ (s)	Notes	Reference
GCaMP6f	0.5 – 0.7	Fast kinetics	Pnevmatikakis 2016
GCaMP6m	1.0 – 1.3	Medium decay	Pnevmatikakis 2016
GCaMP6s	1.5 – 2.0	Slow kinetics	Pnevmatikakis 2016
GCaMP7f	~0.45	Very fast	Rupprecht 2025
GCaMP8f/m	0.2 – 0.3	Modern fast sensors	Rupprecht 2025
GCaMP8s	~0.5	Slow variant	Rupprecht 2025

GCaMP6 Family

Fast (6f): 0.5 s Medium (6m): 1.1 s Slow (6s): 1.8 s

GCaMP7 Family

7f: 0.45 s 7s: 1.0 s 7c: 0.8 s

GCaMP8 Family

8f: 0.25 s 8m: 0.3 s 8s: 0.5 s

Suite2p Spike Deconvolution#

Suite2p’s deconvolution module wraps OASIS, using τ as an internal parameter:contentReference[oaicite:3]{index=3}.
Default values in Suite2p documentation are:

Indicator	Recommended τ (s)	Reference
GCaMP6f	0.7	Suite2p Docs
GCaMP6m	1.0	Suite2p Docs
GCaMP6s	1.25 – 1.5	Suite2p Docs

In in vivo datasets from Dana 2019, these τ values accurately reproduce spike rates for PbN and cortical neurons.
Suite2p applies the same τ≈0.7 s default for red indicators like jRGECO1a (GitHub #233).

CaImAn (CNMF-E Pipeline)#

CaImAn (Giovannucci 2019) uses a similar CNMF model to Suite2p but allows τ tuning within its CNMF-E fitting stage:contentReference[oaicite:4]{index=4}.

Indicator	τ (s)	Implementation Notes
GCaMP6f	0.4 – 0.7	Default for fast decay
GCaMP6m	1.0 – 1.2	Standard for medium kinetics
GCaMP6s	1.5 – 2.0	Slow indicator decay

In most datasets, τ=0.4 s captures fast transients well, while slower indicators require τ > 1 s for stability (Friedrich 2017).

CASCADE (Deep Learning Inference)#

CASCADE (Rupprecht 2021) uses supervised networks trained on simultaneous electrophysiology and calcium data.
τ is not explicitly specified but implicitly learned from training data:contentReference[oaicite:5]{index=5}.

Original 2021 models were trained on GCaMP6 data (τ≈1–2 s).
Updated 2025 versions retrained on GCaMP8f/m/s achieved optimal inference with effective τ ≈ 0.3–0.5 s:contentReference[oaicite:6]{index=6}.

Summary#

Pipeline	Uses τ?	Range (s)	Method	Reference
FOOPSI	Yes	~1.0	Fixed exponential	Vogelstein 2010
OASIS / CNMF	Yes	0.3 – 2.0	AR(1/2) model	Pnevmatikakis 2016
Suite2p	Yes	0.7 – 1.5	OASIS internal	Suite2p Docs
CaImAn	Yes	0.4 – 2.0	CNMF-E fit	Giovannucci 2019
CASCADE	Implicit	0.3 – 2.0	Learned dynamics	Rupprecht 2021

Key Takeaways

τ defines calcium transient decay and sets temporal resolution of spike inference
Optimal τ depends on both indicator kinetics and frame rate
Pipelines like Suite2p and CaImAn require τ tuning per GECI
CASCADE bypasses explicit τ by learning it implicitly
GCaMP8 series are ~3× faster than GCaMP6

Notes#

All τ values summarized here reflect in vivo mammalian calcium imaging (typically ~30 Hz frame rate).
In vitro or temperature-controlled decay times (e.g., 37 °C) can be >10× shorter.
Choosing an incorrect τ biases both spike amplitude and inferred firing rate:contentReference[oaicite:7]{index=7}.

Primary Sources:
Vogelstein 2010 (FOOPSI) ·
Pnevmatikakis 2016 (CNMF) ·
Friedrich 2017 (OASIS) ·
Giovannucci 2019 (CaImAn) ·
Suite2p Docs ·
Dana 2019 (GCaMP7) ·
Rupprecht 2021 & 2025 (CASCADE)