Unifying cognitive aging: From neuromodulation to representation to cognition

doi:10.1016/S0925-2312(00)00256-3

Neurocomputing

Volumes 32–33, June 2000, Pages 879-890

https://doi.org/10.1016/S0925-2312(00)00256-3 Get rights and content

Abstract

An integrative theory relating cognitive aging deficits observed at the behavioral level with age-related deficiency in neuromodulation causing less distinctive cortical representations is tested in a series of neural network simulations. Age-related attenuation of catecholaminergic function is simulated by lowering the mean gain of the processing unit, which subsequently reduces responsivity and raises intra-network activation variability. Age differences in learning rate, asymptotic performance, interference susceptibility, complexity cost, intra- and inter-individual variability, and ability dedifferentiation can all be modeled. Together, the simulations illustrate catecholamine's role in regulating the fidelity of neural information processing and subsequent effects leading to cognitive aging deficits.

Introduction

Declines in basic mechanisms of cognition pervade the aging process. Behaviorally, age differences in processing speed [26], learning rate [16], asymptotic performance [4], [27], interference susceptibility [16], [18], complexity cost [17] and intra-individual [11] as well as inter-individual variability [23] are robust effects. Biologically, brain aging involves neuroanatomical and neurochemical changes [29]. Accounts postulated at either the information-processing level assuming age-related reduction in general processing resources [25] or at the biological level hypothesizing age-related increase in neural noise [34] were proposed as mediating factors for cognitive deficits observed behaviorally. However, thus far data and explanations of cognitive aging deficits have been mostly confined within one (or two) of these levels.

As an attempt to facilitate integration, we offer a cross-level theoretical conjecture aiming at integrating findings of age-related decrements in catecholaminergic function, the functional properties of catecholaminergic modulation, catecholaminergic effects on neural information processing, and various benchmark behavioral manifestations of cognitive aging. Empirical evidences from different levels are reviewed as we unfold and evaluate our conjecture.

Section snippets

Aging, neuromodulation and information processing

The relationship between cognitive aging and age-related deficiency in neuromodulation has recently become an important topic in aging research. Given their roles in modulating prefrontal cortical (PFC) cognitive functions [2], age-related depletion of catecholamine, consisting of dopamine (DA) and its metabolic products, norepinephrine (NE) and epinephrine, is of specific interest. There is consensus of age-related decline in dopaminergic function in the striatum, basal ganglia, and prefrontal

Modeling age-related attenuation of catecholaminergic function

Outside of cognitive aging research, there are quite a few formal models of neuromodulation [7]. However, to our knowledge no research effort has yet been devoted to instantiate formal models that link aging-induced changes in catecholaminergic modulation with a broad class of human cognitive aging phenomena at the behavioral level. In this study, we computationally investigate general principles of neural mechanisms linking catecholaminergic modulation of cortical neuron's responsivity to the

Simulations linking catecholaminergic modulation with behavioral data

The conjectured theoretical path from age-related decrement in catecholaminergic function to higher levels of random variability in the aging brain then to less-distinctive cortical representation and cognitive aging deficits is examined with respect to multiple constraints from benchmark behavioral data. The effects of mean G reduction in accounting for typical cognitive aging phenomena are tested in a series of simulations. To broadly sample different aspects of the behavioral data, the

Conclusions

We have demonstrated with a series of simulations that a single-parameter manipulation (i.e., reducing the mean G of the activation function) accounts for a wide range of cognitive aging phenomena. Regarding unifying data and theories across levels, the simulations explicate a theoretical path from attenuated catecholaminergic function to reduced neural responsivity and increased neural noise to less distinct cortical representation in the aging brain to behavioral manifestations of age-related

References (35)

A.F.T. Arnsten
Catecholamine mechanisms in age-related cognitive decline, Neurobiol. Aging
(1993)
A.F.T. Arnsten
Catecholamine modulation of prefrontal cortical cognitive function
Trends Cog. Sci.
(1998)
L.A. Chiodo et al.
Interactions between dopamine and amino-acid-induced excitation and inhibition in the striatum
Proc. Nat. Acad. Sci. USA
(1986)
P.S. Goldman-Rakic et al.
Regional changes of monoamines in cerebral cortex and subcortical structures of aging rehsus monkeys
Neuroscience
(1981)
R. Kliegl et al.
Time-accuracy functions for determining process and person differences: an application to cognitive aging
Cognitive Psychol.
(1994)
P.G. MacRae et al.
Reaction time and nigrostriatal dopamine function: the effect of age and practice
Brain Res.
(1988)
D. Servan-Schreiber et al.
Dopamine and the mechanisms of cognition: Part I. A neural network model predicting dopamine effects on selective attention
Biol. Psychiatry
(1998)
P.B. Baltes et al.
Integration versus differentiation of fluid/crystallized intelligence in old age
Dev. Psychol.
(1980)
P.B. Baltes et al.
Further testing of the limits of cognitive plasticity: negative age differences in a mnemonic skill are robust
Dev. Psychol.
(1992)
Y. Burnod et al.
Consequences of stochastic release of neurotransmitters of network computation in the central nervous system
Proc. Nat. Acad. Sci. USA
(1989)

J.M. Fellous et al.

Computational models of neuromodulation, Neural Comput.

(1998)

C.E. Finch

Age-related changes in brain catecholamines: a synopsis of findings in C57BL/6J mice and other rodent models

Adv. Exp. Med. Biol.

(1978)

S.L. Foote et al.

Effects of putative neurotransmitters on neuronal activity in monkey auditory cortex

Brain Res.

(1976)

W. Freeman

Non-linear gain mediating cortical stimulus-response relations

J. Biol. Cybernet.

(1979)

J.L. Fozard et al.

Effects of age and frequency of stimulus repetitions on two-choice reaction time

J. Gerontol

(1976)

J.D.E. Gabrieli

Cognitive neuroscience of human memory

Annu. Rev. Psychol.

(1998)

N.A. Hessler et al.

The probability of transmitter release at a mammalian central synapse

Nature

(1993)

Cited by (87)

Greater up-modulation of intra-individual brain signal variability makes a high-load cognitive task more arduous for older adults
2024, NeuroImage
The extent to which brain responses are less distinctive across varying cognitive loads in older adults is referred to as neural dedifferentiation. Moment-to-moment brain signal variability, an emerging indicator, reveals not only the adaptability of an individual's brain as an inter-individual trait, but also the allocation of neural resources within an individual due to ever-changing task demands, thus shedding novel insight into the process of neural dedifferentiation. However, how the modulation of intra-individual brain signal variability reflects behavioral differences related to cognitively demanding tasks remains unclear. In this study, we employed functional near-infrared spectroscopy (fNIRS) imaging to capture the variability of brain signals, which was quantified by the standard deviation, during both the resting state and an n-back task (n = 1, 2, 3) in 57 healthy older adults. Using multivariate Partial Least Squares (PLS) analysis, we found that fNIRS signal variability increased from the resting state to the task and increased with working memory load in older adults. We further confirmed that greater fNIRS signal variability generally supported faster and more stable response time in the 2- and 3-back conditions. However, the intra-individual level analysis showed that the greater the up-modulation in fNIRS signal variability with cognitive loads, the more its accuracy decreases and mean response time increases, suggesting that a greater intra-individual brain signal variability up-modulation may reflect decreased efficiency in neural information processing. Taken together, our findings offer new insights into the nature of brain signal variability, suggesting that inter- and intra-individual brain signal variability may index distinct theoretical constructs.
Functional and structural connectivity correlates of semantic verbal fluency deficits in first-episode psychosis
2024, Journal of Psychiatric Research
Semantic verbal fluency (SVF) impairments are debilitating and present early in the course of psychotic illness. Deficits within frontal, parietal, and temporal brain regions contribute to this deficit, as long-range communication across this functionally integrated network is critical to SVF. This study sought to isolate disruptions in functional and structural connectivity contributing to SVF deficits during first-episode psychosis in the schizophrenia spectrum (FESz).
Thirty-three FESz and 34 matched healthy controls (HC) completed the Animal Naming Task to assess SVF. Magnetoencephalography was recorded during an analogous covert SVF task, and phase-locking value (PLV) used to measure functional connectivity between inferior frontal and temporoparietal structures bilaterally. Diffusion imaging was collected to measure fractional anisotropy (FA) of the arcuate fasciculus, the major tract connecting frontal and temporoparietal language areas.
SVF scores were lower among FESz compared to HC. While PLV and FA did not differ between groups overall, FESz exhibited an absence of the left-lateralized nature of both measures observed in HC. Among FESz, larger right-hemisphere PLV was associated with worse SVF performance (ρ = −0.51) and longer DUP (ρ = −0.50).
In addition to worse SVF, FESz exhibited diminished leftward asymmetry of structural and functional connectivity in fronto-temporoparietal SVF network. The relationship between theta-band hyperconnectivity and poorer performance suggests a disorganized executive network and may reflect dysfunction of frontal cognitive control centers. These findings illustrate an aberrant pattern across the distributed SVF network at disease onset and merit further investigation into development of asymmetrical hemispheric connectivity and its failure among high-risk populations.
Spectral pattern similarity analysis: Tutorial and application in developmental cognitive neuroscience
2022, Developmental Cognitive Neuroscience
The human brain encodes information in neural activation patterns. While standard approaches to analyzing neural data focus on brain (de-)activation (e.g., regarding the location, timing, or magnitude of neural responses), multivariate neural pattern similarity analyses target the informational content represented by neural activity. In adults, a number of representational properties have been identified that are linked to cognitive performance, in particular the stability, distinctiveness, and specificity of neural patterns. However, although growing cognitive abilities across childhood suggest advancements in representational quality, developmental studies still rarely utilize information-based pattern similarity approaches, especially in electroencephalography (EEG) research. Here, we provide a comprehensive methodological introduction and step-by-step tutorial for pattern similarity analysis of spectral (frequency-resolved) EEG data including a publicly available pipeline and sample dataset with data from children and adults. We discuss computation of single-subject pattern similarities and their statistical comparison at the within-person to the between-group level as well as the illustration and interpretation of the results. This tutorial targets both novice and more experienced EEG researchers and aims to facilitate the usage of spectral pattern similarity analyses, making these methodologies more readily accessible for (developmental) cognitive neuroscientists.
A common polymorphism in the dopamine transporter gene predicts working memory performance and in vivo dopamine integrity in aging
2021, NeuroImage
Dopamine (DA) integrity is suggested as a potential cause of individual differences in working memory (WM) performance among older adults. Still, the principal dopaminergic mechanisms giving rise to WM differences remain unspecified. Here, 61 single-nucleotide polymorphisms, located in or adjacent to various dopamine-related genes, were assessed for their links to WM performance in a sample of 1313 adults aged 61–80 years from the Berlin Aging Study II. Least Absolute Shrinkage and Selection Operator (LASSO) regression was conducted to estimate associations between polymorphisms and WM. Rs40184 in the DA transporter gene, SLC6A3, showed allelic group differences in WM, with T-carriers performing better than C homozygotes (p<0.01). This finding was replicated in an independent sample from the Cognition, Brain, and Aging study (COBRA; baseline: n = 181, ages: 64–68 years; 5-year follow up: n = 129). In COBRA, in vivo DA integrity was measured with ¹¹C-raclopride and positron emission tomography. Notably, WM as well as in vivo DA integrity was higher for rs40184 T-carriers at baseline (p<0.05 for WM and caudate and hippocampal D2-receptor availability) and at the 5-year follow-up (p<0.05 for WM and hippocampal D2 availability). Our findings indicate that individual differences in DA transporter function contribute to differences in WM performance in old age, presumably by regulating DA availability.
Neural Dedifferentiation in the Aging Brain
2019, Trends in Cognitive Sciences
Citation Excerpt :
Thus, additional research will be necessary to establish whether evidence of item-level dedifferentiation is manifest in neural measures other than repetition suppression. The computational model of Li and colleagues [9–12] proposes that neural dedifferentiation is an important determinant of cognitive aging. Nonetheless, only a handful of studies have examined whether measures of neural dedifferentiation correlate with cognitive performance, and even fewer have directly examined whether such correlations are moderated by age.
Many cognitive abilities decline with age even in the absence of detectable pathology. Recent evidence indicates that age-related neural dedifferentiation, operationalized in terms of neural selectivity, may contribute to this decline. We review here work exploring the relationship between neural dedifferentiation, cognition, and age. Compelling evidence for age effects on neural selectivity comes from both non-human animal and human research. However, current data suggest that age does not moderate the observed relationships between neural dedifferentiation and cognitive performance. We propose that functionally significant variance in measures of neural dedifferentiation reflects both age-dependent and age-independent factors. We further propose that the effects of age on neural dedifferentiation do not exclusively reflect detrimental consequences of aging.
The impact of the COMT genotype and cognitive demands on facets of intra-subject variability
2019, Brain and Cognition
Intra-Subject Variability (ISV), a potential index of catecholaminergic regulation, is elevated in several disorders linked with altered dopamine function. ISV has typically been defined as reaction time standard deviation. However, the ex-Gaussian and spectral measures capture different aspects and may delineate different underlying sources of ISV; thus reflecting different facets of the construct. We examined the impact of factors associated with dopamine metabolism, namely, Catechol-O-Methyltransferase Val158Met (COMT) genotype and Working Memory (WM) and response-switching on ISV facets in young healthy adults. The Met allele was associated with overall increased variability. The rather exclusive sensitivity of ex-Gaussian tau to frequencies below 0.025 Hz and the quasi-periodic structure of particularly slow responses support the interpretation of tau as low frequency fluctuations of neuronal networks. Sigma, by contrast, may reflect neural noise. Regarding cognitive demands, a WM load-related increase in variability was present for all genotypes and all ISV facets. Contrastingly, ISV facets reacted differently to variations in response-switching as, across genotypes, sigma was elevated for rare target trials whereas tau was elevated for frequent standard trials, particularly for Met homozygotes. Our findings support the significant role of COMT in regulating behavioural ISV with its facetted structure and presumed underlying neural processes.

View all citing articles on Scopus

Shu-Chen Li is a research scientist at the Max Planck Institute for Human Development, Berlin, Germany. She received her Ph.D. from the University of Oklahoma, USA, in December 1994, and was a postdoctoral fellow at McGill University, Montreal, Canada. Her research stretches across behavioral studies of lifespan cognitive development, cognitive neuroscience and computational modeling. She is interested in modeling behavioral data of cognitive performance by computationally implementing neurobiological mechanisms that may have bearings on information-processing mechanism, such as processing speed, inhibition and working memory. She had also published empirical papers on serial-order memory and theoretical articles on using parameter sensitivity analyses as a diagnostic tool for model comparison. She thanks especially Prof. Paul B. Baltes and the Max Planck Institute for sponsoring this project.

Ulman Lindenberger is a research scientist at the Max Planck Institute for Human Development, Berlin. He co-directs several research projects on lifespan cognitive development, and serves on the editorial board of several international journals in the field of cognitive aging. He received his Ph.D. and his habilitation at the Free University of Berlin in 1990 and 1998, respectively. His general research interests are in the field of lifespan cognitive development. His specific interests include the causes and dimensionality of negative age changes in adult cognition, the relationship between age changes in sensory and cognitive functioning, as well as issues regarding experimental and multivariate methodology.

Peter A. Frensch studied electrical engineering, psychology, and philosophy at the Universities of Darmstadt and Trier, Germany, and at Yale University, where he received his M.S., M.Phil., and Ph.D. He worked as an assistant and associate professor in the department of Psychology at the University of Missouri-Columbia and as a senior research scientist at the Max-Plank-Institute for Human Development in Berlin, Germany. Since 1998, he has been a full professor in the department of Psychology at Humboldt-University, Berlin, Germany. He is coordinating editor of the journal Psychological Research. His research interests include human learning, memory, and problem solving.

View full text

Unifying cognitive aging: From neuromodulation to representation to cognition

Abstract

Introduction

Section snippets

Aging, neuromodulation and information processing

Modeling age-related attenuation of catecholaminergic function

Simulations linking catecholaminergic modulation with behavioral data

Conclusions

Catecholamine mechanisms in age-related cognitive decline, Neurobiol. Aging

Trends Cog. Sci.

Proc. Nat. Acad. Sci. USA

Neuroscience

Cognitive Psychol.

Brain Res.

Biol. Psychiatry

Integration versus differentiation of fluid/crystallized intelligence in old age

Dev. Psychol.

Further testing of the limits of cognitive plasticity: negative age differences in a mnemonic skill are robust

Dev. Psychol.

Consequences of stochastic release of neurotransmitters of network computation in the central nervous system

Proc. Nat. Acad. Sci. USA

Computational models of neuromodulation, Neural Comput.

Age-related changes in brain catecholamines: a synopsis of findings in C57BL/6J mice and other rodent models

Adv. Exp. Med. Biol.

Effects of putative neurotransmitters on neuronal activity in monkey auditory cortex

Brain Res.

Non-linear gain mediating cortical stimulus-response relations

J. Biol. Cybernet.

Effects of age and frequency of stimulus repetitions on two-choice reaction time

J. Gerontol

Cognitive neuroscience of human memory

Annu. Rev. Psychol.

The probability of transmitter release at a mammalian central synapse

Nature