Cooperative thalamocortical circuit mechanism for sensory prediction errors

Abstract

The brain functions as a prediction machine, utilizing an internal model of the world to anticipate sensations and the outcomes of our actions. Discrepancies between expected and actual events, referred to as prediction errors, are leveraged to update the internal model and guide our attention towards unexpected events. Despite the importance of prediction error signals for various neural computations across multiple brain regions, surprisingly little is known about the neural circuit mechanisms responsible for their implementation. Here we describe a thalamocortical disinhibitory circuit required for generating sensory prediction errors in mouse primary visual cortex (V1). Using calcium imaging with optogenetic manipulations as mice traverse a familiar virtual environment, we show that violation of animals' predictions by an unexpected visual stimulus preferentially boosts responses of layer 2/3 V1 neurons most selective for that stimulus. Prediction errors specifically amplify the unexpected visual input, rather than representing a non-specific surprise or difference signal about how the visual input deviates from animals' predictions. Selective amplification of unexpected visual input is implemented by a cooperative mechanism requiring thalamic input from the pulvinar, and cortical vasoactive-intestinal-peptide-expressing (VIP) inhibitory interneurons. In response to prediction errors, VIP neurons inhibit a specific subpopulation of somatostatin-expressing (SOM) inhibitory interneurons that gate excitatory pulvinar input to V1, resulting in specific pulvinar-driven response-amplification of the most stimulus-selective neurons in V1. Therefore, the brain prioritizes unpredicted sensory information by selectively increasing the salience of unpredicted sensory features through the synergistic interaction of thalamic input and neocortical disinhibitory circuits.

Experience-dependent predictions of feedforward and contextual information in mouse visual cortex

“…A reduction of responses to expected and frequently presented stimuli is a known and robust effect 31–36 , but it is debated whether this primarily affects the most selective or the least selective neurons, causing dampening or sharpening of the responses, respectively. We therefore tested how the stimulus selectivity (lifetime sparseness, which measures the selectivity of a neuron’s response to different stimuli) for non-occluded images changed after familiarization and found that it increased in both L2/3 and L5 (Figure 3F and 4F, LMEM: L2/3 naive versus expert P < 0.001 and L5 naive versus expert P < 0.001).…”

Selective engagement of prefrontal VIP neurons in reversal learning

“…As potent regulators, VIP neurons might specialize in monitoring significant changes in environmental contingencies and issuing reward-related error correction signals, thereby facilitating rapid behavioral adjustments. VIP neurons in sensory cortical areas show enhanced responsiveness to novel or unexpected sensory stimuli 35,36 , and they dynamically modulate other neurons' sensory responses according to factors such as locomotion, behavioral strategy, and arousal [37][38][39][40] . These findings, together with our own, suggest that cortical VIP neurons may specialize in a supervisory role, wherein they oversee notable deviations from established internal and external states and intervene in ongoing cortical processes as necessary.…”

How predictability and individual alpha frequency shape memory: insights from an event-related potential investigation

“…This suggests that prediction errors may drive the formation of new memories; a proposal that is supported by behavioural research demonstrating enhanced memory for incongruent stimuli such as reward prediction errors (Jang et al, 2019; Rouhani et al, 2018), surprising images (Greve et al, 2017; Michelon et al, 2003), bizarre sentences (Worthen & Roark, 2002), novel words (Habib et al, 2003), and unpredictable, implausible nouns (however, see Haeuser & Kray, 2022 for a discussion of this effect as being driven by the implausibility of the word, as opposed to the predictability itself). To further support the suggestion that the brain prioritises unpredictable information, Furutachi and colleagues (2023) observed amplified sensory representations in the visual cortex of mice for prediction error-eliciting stimuli. As increased brain activity at the time of encoding has been positively correlated with memory outcomes (Kim, 2011), the amplification of error-inducing data may promote stronger memory performance.…”