Two Routes to Interoceptive Prioritisation: Dissociating Likelihood Precision and Preference Reweighting in Active Inference

St John Grimbly; Nicolas Kuske; Evert A. Boonstra; Charel van Hoof; Rowan Hodson; Benjamin Rosman; Ryan Smith; Mark Solms

cs.AIOct 13, 2026

Two Routes to Interoceptive Prioritisation: Dissociating Likelihood Precision and Preference Reweighting in Active Inference

Authors: St John Grimbly, Nicolas Kuske, Evert A. Boonstra, Charel van Hoof, Rowan Hodson, Benjamin Rosman, Ryan Smith, Mark Solms,

Organizations: Dept. of Mathematics & Applied Mathematics, Univ. of Cape Town, South Africa · Neuroscience Institute, Univ. of Cape Town, South Africa · Delft University of Technology, Dept. of Cognitive Robotics · Laureate Institute for Brain Research, Tulsa, OK, USA ·

Abstract

An embodied agent must prioritise among competing bodily needs. In active inference, it can respond to an urgent need by increasing the precision of information about that need or by increasing its preference for outcomes that relieve it. We compare these two interventions, separately and together, in a foraging gridworld with three lethal needs, and ask how each changes survival, perception and the body model the agent learns. Precision sharpens interoceptive posteriors and improves survival. Preference reweighting keeps bodily states further from their lethal boundaries and changes the states from which the agent learns, so its learned model covers fewer need levels, with lower quality across all levels but slightly higher quality on the levels it visits. An experiment on 72 further layouts does not establish the predicted preference benefit on difficult layouts, but both interventions improve average survival there and their interaction is positive. Survival alone therefore gives an incomplete account of how a model prioritises needs and learns about its body.

Figures & tables

Figure 1: The foraging task (illustrative, with layouts varying): one food and one water tile on a

6\times 6

grid. The needs panel (right) tracks three interoceptive levels and their lethal floors. Hunger and thirst decline over time. Drinking restores thirst but lowers oxygen, which recovers only on land. The best priority therefore depends on the current body state.

Figure 2: The perception–action loop and the two implemented mechanisms. Precision (

\kappa

) sharpens the likelihood

A

, which feeds inference and planning and governs the body’s emission. Preference reweighting changes

C

, which enters only the planner’s pragmatic term. The lower boxes show principal internal effects on the original layouts, not effects exclusive to either intervention.

Figure 3: Survival and the internal dissociation. (a) Condition means and layout-bootstrap

95\%

intervals across eleven layouts and eight seeds. (b) Layout-level signal-to-error (

|\mathrm{mean}/\mathrm{s.e.}|

). Bar length is descriptive. Solid bars have

95\%

intervals excluding zero (Tables 1 , 3 ).

Contrast	$\Delta$ survival	95% CI	Holm $p$
$\kappa$ -routing vs uniform baseline	$+0.27$	$[0.12,\ 0.44]$	$0.0176$
dynamic- $C$ vs uniform baseline	$+0.11$	$[-0.06,\ 0.27]$	$0.248$
$\kappa$ -routing+dynamic- $C$ vs uniform baseline	$+0.38$	$[0.18,\ 0.56]$	$0.0176$

Table 1: Survival contrasts on the eleven original layouts with eight paired seeds. Intervals are paired layout bootstraps. The

p

values use exact layout-level sign flips with Holm adjustment across the three contrasts.

Comparison	Effect	95% CI	$p$
Preference, difficult layouts (primary)	$-0.008$	$[-0.025,\,+0.016]$	$0.5793$
Precision, all layouts (secondary)	$+0.111$	$[+0.076,\,+0.147]$	$<0.0001$
Preference, all layouts (secondary)	$+0.080$	$[+0.046,\,+0.118]$	$<0.0001$
Interaction, all layouts (secondary)	$+0.114$	$[+0.049,\,+0.182]$	$0.002$

Table 2: Survival effects in the additional experiment. The primary test uses the 24 layouts calibrated as difficult. Secondary effects average over all 72 layouts. Intervals resample layouts. Secondary

p

values are Holm-adjusted jointly with the continuous difficulty test.

Metric (vs. baseline)	$\kappa$ -routing	dynamic- $C$	$\kappa$ -routing+dynamic- $C$
Posterior entropy ( $10^{-3}$ )	$-1.37\,[\text{-}1.95,\text{-}0.84]$	$-0.26\,[\text{-}0.83,0.28]$	$-1.18\,[\text{-}1.77,\text{-}0.58]$
Visited-state entropy	$+0.00\,[\text{-}0.09,0.07]$	$-0.05\,[\text{-}0.14,0.02]$	$-0.15\,[\text{-}0.24,\text{-}0.08]$
Minimum headroom	$+0.01\,[\text{-}0.02,0.04]$	$+0.05\,[0.01,0.09]$	$+0.11\,[0.07,0.15]$

Table 3: Internal contrasts against baseline, shown as mean [layout-bootstrap

95\%

CI]. Entropies use

10^{-3}

nats (posterior) and nats (visited state). Headroom is a fraction.

Active Inference

Active inference studies how agents perceive, learn, and act by updating generative models and minimizing expected free energy.

See all 42 papers

Interoceptive Attention as Dynamic Homeostatic Prioritization in a Foraging Agent
Aug 4, 2026St John Grimbly, Nicolas Kuske, Evert A. Boonstra +7Adaptive Inference Active Inference
Dept. of Mathematics & Applied Mathematics, Univ. of Cape Town, South Africa · Neuroscience Institute, Univ. of Cape Town, South Africa · MIND Institute and CSAM, University of the Witwatersrand, South Africa +3
Belief-Based Maximum Occupancy Principle and Active Inference
Sep 30, 2026Manolis Mylonas, Rubén Moreno BoteBelief-Space Planning Intrinsic Reward Methods
Center for Brain and Cognition & Department of Engineering Universitat Pompeu Fabra Barcelona, Spain
Asking Smarter Questions: Targeted Active Preference Learning for Underspecified Rewards
Oct 15, 2026Helena Merker, Nick Walker, Andreea BobuActive Preference Learning Preference Learning
Massachusetts Institute of Technology

Aug 4, 2026J/K move · Enter open · S save

Interoceptive Attention as Dynamic Homeostatic Prioritization in a Foraging Agent

St John Grimbly, Nicolas Kuske, Evert A. Boonstra, Bruce A. Bassett+6

Dept. of Mathematics & Applied Mathematics, Univ. of Cape Town, South Africa · Neuroscience Institute, Univ. of Cape Town, South Africa · MIND Institute and CSAM, University of the Witwatersrand, South Africa+3

Adaptive Inference Active Inference

Open paper AskPDF arXiv

Biological systems must regulate competing needs under limited perceptual bandwidth, where sharpening one estimate costs the capacity to sharpen the others. Any fixed-budget system therefore has to decide where to allocate its perceptual precision. We study this in a foraging agent that must keep several bodily needs satisfied to survive, modelled with active inference. At each step it reads its own body-state beliefs, identifies the most-needed channel, and reallocates a fixed budget of interoceptive precision toward it, so that the same precision-shaped likelihood feeds both belief update and planning. In AffectWorld, a four-channel foraging gridworld, this selective allocation more than doubles learning-phase survival at matched budget against a uniform-precision agent (

0.414

0.199

across 11 layouts,

n{=}32

seeds each, paired cluster-bootstrap

p \leq 10^{-4}

). Two further results sharpen the mechanism. The benefit runs through planning as well as perception, since denying the shaped likelihood to the planner alone removes about half of it. It is also need-aligned, since aiming precision at the least-needed channel does worse than spreading it evenly. The attended channel additionally learns its own dynamics about twice as fast, and stays ahead even at matched observation count, a behavioural trace of the same precision routing, visible in learning speed, not survival.

Abstract

Active Inference

Explore similar work

Interoceptive Attention as Dynamic Homeostatic Prioritization in a Foraging Agent

Belief-Based Maximum Occupancy Principle and Active Inference

Asking Smarter Questions: Targeted Active Preference Learning for Underspecified Rewards