Mechanisms underlying working memory for novel information

doi:10.1016/j.tics.2006.09.005

Trends in Cognitive Sciences

Volume 10, Issue 11, November 2006, Pages 487-493

https://doi.org/10.1016/j.tics.2006.09.005 Get rights and content

In this Opinion article we describe a theory that the brain mechanisms underlying working memory for novel information include a buffer in parahippocampal cortices. Computational modeling indicates that mechanisms for maintaining novel information in working memory could differ from mechanisms for maintaining familiar information. Electrophysiological data suggest that the buffer for novel information depends on acetylcholine. Acetylcholine activates single-cell mechanisms that underlie persistent spiking of neurons in the absence of synaptic transmission, allowing maintenance of information without prior synaptic modification. fMRI studies and lesion studies suggest that parahippocampal regions mediate working memory for novel stimuli, and the effects of cholinergic blockade impair this function. These intrinsic mechanisms in parahippocampal cortices provide an important alternative to theories of working memory based on recurrent synaptic excitation.

Introduction

Research has demonstrated the existence of multiple memory systems, including working memory, which is defined as a limited capacity system for the temporary storage and manipulation of information for cognitive tasks 1, 2. Many initial fMRI studies of working memory in humans used highly familiar stimuli such as letters and words, and focused primarily on a system that includes prefrontal cortex and parietal cortex 1, 2, 3. However, electrophysiological and lesion studies in animals and recent imaging studies suggest that temporal lobe structures play a crucial role in working memory 4, 5, 6. Here we present the theory that working memory for novel information differs from working memory for familiar information, proposing that novel stimuli require additional cellular mechanisms within the entorhinal and perirhinal cortex, whereas prefrontal and parietal systems are sufficient for maintaining familiar stimuli in working memory. We are not proposing a double dissociation between the systems, but instead we suggest that the prefrontal–parietal system alone is insufficient for maintaining information that has no prior representation in the brain. We hypothesize that cholinergic activation of these single neuron mechanisms results in persistent spiking activity without excitatory synaptic transmission between neurons (Figure 1 and Box 1). This theory provides an alternative to many physiological models of working memory that use recurrent excitation between neurons to maintain persistent spiking [7]. Those other models, based on recurrent excitation, can only maintain information consistent with previously formed representations (i.e. familiar information).

In this article, we evaluate experimental data 6, 8, 9, 10 and computational modeling 11, 12, 13 that support the proposal that working memory for novel stimuli requires additional cellular mechanisms activated by acetylcholine in the entorhinal cortex and other parahippocampal cortices and that differ substantially from the mechanisms required for working memory for familiar stimuli. In each section we first evaluate evidence for the role of this cellular mechanism in the active maintenance of novel stimuli for working memory, and, if evidence is available, we consider a role for this mechanism in encoding long-term memories.

Section snippets

Working memory

Lesion studies suggest an important role for parahippocampal regions in working memory for novel stimuli. Although cortical regions including prefrontal cortex and parietal cortex show activity during working memory for novel stimuli [14], these regions do not seem to be sufficient to perform working memory for novel stimuli at normal levels when parahippocampal regions are lesioned, although they are sufficient to maintain normal working memory for familiar stimuli. Many studies test working

Working memory

Functional neuroimaging studies support the idea that working memory for novel stimuli requires the additional recruitment of parahippocampal regions. Early functional neuroimaging studies of working memory emphasized the role of prefrontal and parietal cortices, and most studies were carried out using highly familiar stimuli, including letters, words, simple objects and spatial locations [26]. Surprisingly, these early fMRI studies of working memory did not report activity within

Working memory

Studies of the behavioral effects of drugs that block muscarinic acetylcholine receptors support a role for cholinergic modulation in the entorhinal cortex in the short-term retention of novel memories. Systemic injections of muscarinic acetylcholine blockers such as scopolamine impair DMS performance in monkeys at delays on the order of several seconds [31], and impair the arm choice behavior of rats in an eight-arm radial maze when there is a delay between the individual choices [32]. In

Cellular physiology

The proposal of a working memory system for novel stimuli presented here was initially motivated by intracellular single-cell recordings of memory mechanisms 8, 47. Recordings in the laboratory of Angel Alonso demonstrated that single neurons in layer II of slice preparations of the entorhinal cortex, isolated from other neurons by pharmacological blockade of excitatory and inhibitory synaptic transmission, can maintain memory for prior input in the form of persistent spiking activity (Figure 3

Recording of spiking activity during behavior

The theory that the Alonso current in the entorhinal cortex contributes to the maintenance of working memory is supported by recordings of the spiking activity of single neurons in animals performing delayed matching tasks. Sustained spiking activity during delay periods has been shown for neurons in parahippocampal cortices, including the entorhinal cortex, in rats performing a continuous DNMS task with odors [10] and in monkeys during performance of DMS tasks with complex visual stimuli [30].

Summary and future directions

The experimental data and computational modeling reviewed here support the hypothesis that intrinsic single neuron mechanisms in parahippocampal cortices mediate working memory for novel stimuli. Studies using techniques at several different levels support this hypothesis. However, further work at the behavioral level is needed to test whether blockade of cholinergic receptors will selectively impair active maintenance of novel but not familiar stimuli in the same task in humans, possibly by

Acknowledgements

This work was supported by NINDS 41636, NIMH 60013, NIMH 61492, NIMH 60450, NSF Science of Learning Center SBE0354378 and NIDA 16454 as part of the joint NSF/NIH program for Collaborative Research in Computational Neuroscience (CRCNS).

References (60)

A. Baddeley
The episodic buffer: a new component of working memory?
Trends Cogn. Sci.
(2000)
J.M. Fuster
Memory networks in the prefrontal cortex
Prog. Brain Res.
(2000)
C. Ranganath et al.
Directing the mind's eye: prefrontal, inferior and medial temporal mechanisms for visual working memory
Curr. Opin. Neurobiol.
(2005)
W.F. Asaad
Neural activity in the primate prefrontal cortex during associative learning
Neuron
(1998)
C. Ranganath et al.
Doubts about double dissociations between short- and long-term memory
Trends Cogn. Sci.
(2005)
R.A. Koene
Modeling goal-directed spatial navigation in the rat based on physiological data from the hippocampal formation
Neural Netw.
(2003)
O. Jensen et al.
Hippocampal sequence-encoding driven by a cortical multi-item working memory buffer
Trends Neurosci.
(2005)
T.S. Braver
A parametric study of prefrontal cortex involvement in human working memory
Neuroimage
(1997)
C. Ranganath et al.
Medial temporal lobe activity associated with active maintenance of novel information
Neuron
(2001)
D.M. Penetar et al.
Effects of cholinergic drugs on delayed match to sample performance of Rhesus monkeys
Pharmacol. Biochem. Behav.
(1983)

J.J. Bolhuis

Effects of scopolamine on performance of rats in a delayed-response radial maze task

Physiol. Behav.

(1988)

A. Green

Muscarinic and nicotinic receptor modulation of object and spatial n-back working memory in humans

Pharmacol. Biochem. Behav.

(2005)

W.W. Beatty

Patterns of memory failure after scopolamine treatment: implications for cholinergic hypotheses of dementia

Behav. Neural Biol.

(1986)

P. Broks

Modelling dementia: effects of scopolamine on memory and attention

Neuropsychologia

(1988)

T.G. Aigner

Comparison of the effects of scopolamine administered before and after acquisition in a test of visual recognition memory in monkeys

Behav. Neural Biol.

(1991)

J.Z. Xiang et al.

Neuronal responses related to long-term recognition memory processes in prefrontal cortex

Neuron

(2004)

E.C. Warburton

Cholinergic neurotransmission is essential for perirhinal cortical plasticity and recognition memory

Neuron

(2003)

E. Fransen

Mechanism of graded persistent cellular activity of entorhinal cortex layer v neurons

Neuron

(2006)

M.W. Howard et al.

A distributed representation of temporal context

J. Math. Psychol.

(2002)

H. Lee

Phase locking of single neuron activity to theta oscillations during working memory in monkey extrastriate visual cortex

Neuron

(2005)

H. Eichenbaum et al.

From Conditioning to Conscious Recollection: Memory Systems of the Brain

(2001)

C.E. Stern

Medial temporal and prefrontal contributions to working memory tasks with novel and familiar stimuli

Hippocampus

(2001)

K. Schon

Scopolamine reduces persistent activity related to long-term encoding in the parahippocampal gyrus during delayed matching in humans

J. Neurosci.

(2005)

D. Durstewitz

Neurocomputational models of working memory

Nat. Neurosci.

(2000)

R. Klink et al.

Muscarinic modulation of the oscillatory and repetitive firing properties of entorhinal cortex layer II neurons

J. Neurophysiol.

(1997)

A.V. Egorov

Graded persistent activity in entorhinal cortex neurons

Nature

(2002)

B.J. Young

Memory representation within the parahippocampal region

J. Neurosci.

(1997)

J.E. Lisman et al.

Storage of 7 ± 2 short-term memories in oscillatory subcycles

Science

(1995)

O. Jensen et al.

Novel lists of 7 ± 2 known items can be reliably stored in an oscillatory short-term memory network: interaction with long-term memory

Learn. Mem.

(1996)

E. Fransen

Simulations of the role of the muscarinic-activated calcium-sensitive nonspecific cation current INCM in entorhinal neuronal activity during delayed matching tasks

J. Neurosci.

(2002)

Cited by (203)

Bioinspired membrane learnable spiking neural network for autonomous vehicle sensors fault diagnosis under open environments
2023, Reliability Engineering and System Safety
Autonomous vehicles have successfully driven autonomously on urban roads, relying on numerous sensors for environmental perception and vehicle control. However, the abnormality and degradation of sensors will make vehicles face serious safety risks. Therefore, autonomous vehicles must have complete sensor fault diagnosis systems to detect anomalies and avoid accidents. Therefore, this paper explores brain-inspired spiking neural networks (SNN) for sensor fault diagnosis. Specifically, this paper proposes a brain-inspired membrane learnable residual spiking neural network (MLR-SNN) for sensor fault and health index prediction. SNN accurately simulates the dynamic mechanism of biological neurons and exhibits excellent spatiotemporal information processing potential and low power consumption while being highly biologically credible. Based on the convolution topology, this study designs a spike-residual-based SNN framework that optimizes the gradient transfer efficiency to enable deep-level spiking information encoding. In addition, membrane-learnable mechanisms are introduced to simulate the differences of neuronal membrane-related parameters in brains, which can better characterize the dynamics of neurons. The proposed MLR-SNN is validated on actual autonomous vehicle sensor datasets. Experimental results show that MLR-SNN with neural dynamics mechanism has excellent performance, and it can accurately predict fault mode and health index from multivariate sensor data under open environments.
Distinct hippocampal-prefrontal neural assemblies coordinate memory encoding, maintenance, and recall
2023, Current Biology
Short-term memory enables incorporation of recent experience into subsequent decision-making. This processing recruits both the prefrontal cortex and hippocampus, where neurons encode task cues, rules, and outcomes. However, precisely which information is carried when, and by which neurons, remains unclear. Using population decoding of activity in rat medial prefrontal cortex (mPFC) and dorsal hippocampal CA1, we confirm that mPFC populations lead in maintaining sample information across delays of an operant non-match to sample task, despite individual neurons firing only transiently. During sample encoding, distinct mPFC subpopulations joined distributed CA1-mPFC cell assemblies hallmarked by 4–5 Hz rhythmic modulation; CA1-mPFC assemblies re-emerged during choice episodes but were not 4–5 Hz modulated. Delay-dependent errors arose when attenuated rhythmic assembly activity heralded collapse of sustained mPFC encoding. Our results map component processes of memory-guided decisions onto heterogeneous CA1-mPFC subpopulations and the dynamics of physiologically distinct, distributed cell assemblies.
Persistent neuronal firing in the medial temporal lobe supports performance and workload of visual working memory in humans
2022, NeuroImage
The involvement of the medial temporal lobe (MTL) in working memory is controversially discussed. Recent findings suggest that persistent neural firing in the hippocampus during maintenance in verbal working memory is associated with workload. Here, we recorded single neuron firing in 13 epilepsy patients (7 male) while they performed a visual working memory task. The number of colored squares in the stimulus set determined the workload of the trial. Performance was almost perfect for low workload (1 and 2 squares) and dropped at high workload (4 and 6 squares), suggesting that high workload exceeded working memory capacity. We identified maintenance neurons in MTL neurons that showed persistent firing during the maintenance period. More maintenance neurons were found in the hippocampus for trials with correct compared to incorrect performance. Maintenance neurons increased and decreased firing in the hippocampus and increased firing in the entorhinal cortex for high compared to low workload. Population firing predicted workload particularly during the maintenance period. Prediction accuracy of workload based on single-trial activity during maintenance was strongest for neurons in the entorhinal cortex and hippocampus. The data suggest that persistent neural firing in the MTL reflects a domain-general process of maintenance supporting performance and workload of multiple items in working memory below and beyond working memory capacity. Persistent neural firing during maintenance in the entorhinal cortex may be associated with its preference to process visual-spatial arrays.
Role of prediction error and the cholinergic system on memory reconsolidation processes in mice
2021, Neurobiology of Learning and Memory
Citation Excerpt :
These receptors can be divided in two broad families: nicotinic (nAChRs, ligan-gated ion channels) (Albuquerque, Pereira, Alkondon, & Rogers, 2009) and muscarinic (mAChRs, a seven -transmembrane G-protein-coupled receptor) (Bubser, Byun, Wood, & Jones, 2012). Previous studies showed that the cholinergic system facilitates cue detection and attentional performance (Hasselmo & Sarter, 2011; Hasselmo & Stern, 2006). A blockade of the cholinergic system with pre-training administration of specific antagonists leads to changes in attentional processes and encoding; possibly altering acquisition of the novel information and causing subsequent memory impairment.
The ability to make predictions based on stored information is a general coding strategy. A prediction error (PE) is a mismatch between expected and current events. Our memories, like ourselves, are subject to change. Thus, an acquired memory can become active and update its content or strength by a labilization-reconsolidation process. Within the reconsolidation framework, PE drives the updating of consolidated memories. In the past our lab has made key progresses showing that a blockade in the central cholinergic system during reconsolidation can cause memory impairment, while reinforcement of cholinergic activity enhances it. In the present work we determined that PE is a necessary condition for memory to reconsolidate in an inhibitory avoidance task using both male and female mice. Depending on the intensity of the unconditioned stimulus (US) used during training, a negative (higher US intensity) or positive (lower US intensity/no US) PE on a retrieval session modified the behavioral response on a subsequent testing session. Furthermore, we demonstrated that the cholinergic system modulates memory reconsolidation only when PE is detected. In this scenario administration of oxotremorine, scopolamine or nicotine after memory reactivation either enhanced or impaired memory reconsolidation in a sex-specific manner.
The Unexplored Territory of Neural Models: Potential Guides for Exploring the Function of Metabotropic Neuromodulation
2021, Neuroscience
The space of possible neural models is enormous and under-explored. Single cell computational neuroscience models account for a range of dynamical properties of membrane potential, but typically do not address network function. In contrast, most models focused on network function address the dimensions of excitatory weight matrices and firing thresholds without addressing the complexities of metabotropic receptor effects on intrinsic properties. There are many under-explored dimensions of neural parameter space, and the field needs a framework for representing what has been explored and what has not. Possible frameworks include maps of parameter spaces, or efforts to categorize the fundamental elements and molecules of neural circuit function. Here we review dimensions that are under-explored in network models that include the metabotropic modulation of synaptic plasticity and presynaptic inhibition, spike frequency adaptation due to calcium-dependent potassium currents, and afterdepolarization due to calcium-sensitive non-specific cation currents and hyperpolarization activated cation currents. Neuroscience research should more effectively explore possible functional models incorporating under-explored dimensions of neural function.
Delay-dependent cholinergic modulation of visual short-term memory in rhesus macaques
2021, Behavioural Brain Research
Cholinergic neuromodulation is known to play a key role in visual working memory (VWM) – keeping relevant stimulus representations available for cognitive processes for short time periods (up to a few minutes). Despite the growing body of evidence on how the neural and cognitive mechanisms of VWM dynamically change over retention time, there is mixed evidence available on cholinergic effects as a function of VWM delay period in non-human primates. Using the delayed matching to sample VWM task in rhesus macaques (N = 6), we aimed to characterize VWM maintenance in terms of performance changes as a function of delay duration (across a wide range of delays from 1 to 76 s). Then, we studied how cholinergic neuromodulation influences VWM maintenance using the muscarinic receptor antagonist scopolamine administered alone as transient amnestic treatment, and in combination with two doses of the acetylcholinesterase inhibitor donepezil, a widely used Alzheimer’s medication probing for the reversal of scopolamine-induced impairments. Results indicate that scopolamine-induced impairments of VWM maintenance are delay-dependent and specifically affect the 15−33 s time range, suggesting that scopolamine worsens the normal decay of VWM with the passage of time. Donepezil partially rescued the observed scopolamine-induced impairments of VWM performance. These results provide strong behavioral evidence for the role of increased cholinergic tone and muscarinic neuromodulation in the maintenance of VWM beyond a few seconds, in line with our current knowledge on the role of muscarinic acetylcholine receptors in sustained neural activity during VWM delay periods.

View all citing articles on Scopus

View full text

Trends in Cognitive Sciences

OpinionMechanisms underlying working memory for novel information

Introduction

Section snippets

Working memory

Working memory

Working memory

Cellular physiology

Recording of spiking activity during behavior

Summary and future directions

Acknowledgements

Trends Cogn. Sci.

Prog. Brain Res.

Curr. Opin. Neurobiol.

Neuron

Trends Cogn. Sci.

Neural Netw.

Trends Neurosci.

Neuroimage

Neuron

Pharmacol. Biochem. Behav.

Physiol. Behav.

Pharmacol. Biochem. Behav.

Behav. Neural Biol.

Neuropsychologia

Behav. Neural Biol.

Neuron

Neuron

Neuron

J. Math. Psychol.

Neuron

From Conditioning to Conscious Recollection: Memory Systems of the Brain

Medial temporal and prefrontal contributions to working memory tasks with novel and familiar stimuli

Hippocampus

Scopolamine reduces persistent activity related to long-term encoding in the parahippocampal gyrus during delayed matching in humans

J. Neurosci.

Neurocomputational models of working memory

Nat. Neurosci.

Muscarinic modulation of the oscillatory and repetitive firing properties of entorhinal cortex layer II neurons

J. Neurophysiol.

Graded persistent activity in entorhinal cortex neurons

Nature

Memory representation within the parahippocampal region

J. Neurosci.

Storage of 7 ± 2 short-term memories in oscillatory subcycles

Science

Novel lists of 7 ± 2 known items can be reliably stored in an oscillatory short-term memory network: interaction with long-term memory

Learn. Mem.

Simulations of the role of the muscarinic-activated calcium-sensitive nonspecific cation current INCM in entorhinal neuronal activity during delayed matching tasks

J. Neurosci.

Opinion
Mechanisms underlying working memory for novel information