Recurrent neural networks with explicit representation of dynamic latent variables can mimic behavioral patterns in a physical inference task

Personalized brain models link cognitive decline progression to underlying synaptic and connectivity degeneration

Cognitive decline is a condition affecting almost one sixth of the elder population and is widely regarded as one of the first manifestations of Alzheimer's disease. Despite the extensive body of knowledge on the condition, there is no clear consensus on the structural defects and neurodegeneration processes determining cognitive decline evolution. Here, we introduce a Brain Network Model (BNM) simulating the effects of neurodegeneration on neural activity during cognitive processing. The model incorporates two key parameters accounting for distinct pathological mechanisms: synaptic degeneration, primarily leading to hyperexcitation, and brain disconnection. Through parameter optimization, we successfully replicated individual electroencephalography (EEG) responses recorded during task execution from 145 participants spanning different stages of cognitive decline. The cohort included healthy controls, patients with subjective cognitive decline (SCD), and those with mild cognitive impairment (MCI) of the Alzheimer type. Through model inversion, we generated personalized BNMs for each participant based on individual EEG recordings. These models revealed distinct network configurations corresponding to the patient's cognitive condition, with virtual neurodegeneration levels directly proportional to the severity of cognitive decline. Strikingly, the model uncovered a neurodegeneration-driven phase transition leading to two distinct regimes of neural activity underlying task execution. On either side of this phase transition, increasing synaptic degeneration induced changes in neural activity that closely mirrored experimental observations across cognitive decline stages. This enabled the model to directly link synaptic degeneration and hyperexcitation to cognitive decline severity. Furthermore, the model pinpointed posterior cingulum fiber degeneration as the structural driver of this phase transition. Our findings highlight the potential of BNMs to account for the evolution of neural activity across stages of cognitive decline while elucidating the underlying neurodegenerative mechanisms. This approach provides a novel framework for understanding how structural and functional brain alterations contribute to cognitive deterioration along the Alzheimer's continuum.

Advanced Artificial Intelligence Technologies Transforming Contemporary Pharmaceutical Research

One area of study within machine learning and artificial intelligence (AI) seeks to create computer programs with intelligence that can mimic human focal processes in order to produce results. This technique includes data collection, effective data usage system development, conclusion illustration, and arrangements. Analysis algorithms that are learning to mimic human cognitive activities are the most widespread application of AI. Artificial intelligence (AI) studies have proliferated, and the field is quickly beginning to understand its potential impact on medical services and investigation. This review delves deeper into the pros and cons of AI across the healthcare and pharmaceutical research industries. Research and review articles published throughout the last few years were selected from PubMed, Google Scholar, and Science Direct, using search terms like 'artificial intelligence', 'drug discovery', 'pharmacy research', 'clinical trial', etc. This article provides a comprehensive overview of how artificial intelligence (AI) is being used to diagnose diseases, treat patients digitally, find new drugs, and predict when outbreaks or pandemics may occur. In artificial intelligence, neural networks and deep learning are some of the most popular tools; in clinical research, Bayesian non-parametric approaches hold promise for better results, while smartphones and the processing of natural languages are employed in recognizing patients and trial monitoring. Seasonal flu, Ebola, Zika, COVID-19, tuberculosis, and outbreak predictions were made using deep computation and artificial intelligence. The academic world is hopeful that AI development will lead to more efficient and less expensive medical and pharmaceutical investigations and better public services.

MARBLE: interpretable representations of neural population dynamics using geometric deep learning

The dynamics of neuron populations commonly evolve on low-dimensional manifolds. Thus, we need methods that learn the dynamical processes over neural manifolds to infer interpretable and consistent latent representations. We introduce a representation learning method, MARBLE, which decomposes on-manifold dynamics into local flow fields and maps them into a common latent space using unsupervised geometric deep learning. In simulated nonlinear dynamical systems, recurrent neural networks and experimental single-neuron recordings from primates and rodents, we discover emergent low-dimensional latent representations that parametrize high-dimensional neural dynamics during gain modulation, decision-making and changes in the internal state. These representations are consistent across neural networks and animals, enabling the robust comparison of cognitive computations. Extensive benchmarking demonstrates state-of-the-art within- and across-animal decoding accuracy of MARBLE compared to current representation learning approaches, with minimal user input. Our results suggest that a manifold structure provides a powerful inductive bias to develop decoding algorithms and assimilate data across experiments.

Neural circuit mechanisms to transform cerebellar population dynamics for motor control in monkeys

We exploit identification of neuron types during extracellular recording to demonstrate how the cerebellar cortex's well-established architecture transforms inputs into outputs. During smooth pursuit eye movements, the floccular complex performs distinct input-output transformations of temporal dynamics and directional response properties. The responses of different interneuron types localize the circuit mechanisms of each transformation. Mossy fibers and unipolar brush cells emphasize eye position dynamics uniformly across the cardinal axes; Purkinje cells and molecular layer interneurons code eye velocity along directionally biased axes; Golgi cells show unmodulated firing. Differential directional response properties of different neuron types localize the directional input-output transformation to the last-order inputs to Purkinje cells. Differential temporal dynamics pinpoint the site of the temporal input-output transformation to granule cells. Specific granule cell population dynamics allow the temporal transformations required in the area we study and generalize to many temporal transformations, providing a complete framework to understand cerebellar circuit computation.

Impact statement

We dissect the circuit computations performed by the floccular complex of the cerebellum during an exemplar sensory-motor behavior, taking advantage of knowledge of the circuit architecture, existence of discrete neuron types, and a newfound ability to identify neuron types from extracellular recordings. Our results describe the contributions of the major neuron types to the cerebellar input-output computations, identify the population dynamics needed in granule cells to support those computations, and to create a basis set to enable temporally-specific motor behavior and motor learning.

A language model of problem solving in humans and macaque monkeys

Human intelligence is characterized by the remarkable ability to solve complex problems by planning a sequence of actions that takes us from an initial state to a desired goal state. Quantifying and comparing problem-solving capabilities across species and finding their evolutionary roots are critical for understanding how the brain carries out this intricate process. We introduce the Language of Problem Solving (LoPS) model as a novel quantitative framework that investigates the structure of problem-solving behavior through a language model. We applied the model to an adapted classic Pac-Man game as a cross-species behavioral paradigm to test both humans and macaque monkeys. The LoPS model extracted the latent structure, or grammar, embedded in the agents' gameplay, revealing the non-Markovian temporal dependency structure of their problem-solving behavior and the hierarchical structures of problem solving in both species. The complexity of LoPS grammar correlated with individuals' game performance and reflected the difference in problem-solving capacity between humans and monkeys. Both species evolved their LoPS grammars during learning, progressing from simpler to more complex ones, suggesting that the structure of problem solving is not fixed but evolves to support more sophisticated and efficient problem solving. Our study provides insights into how humans and monkeys break down problem solving into compositional units and navigate complex tasks, deepening our understanding of human intelligence and its evolution and establishing a foundation for future investigations of the neural mechanisms of problem solving.

Dynamic tracking of objects in the macaque dorsomedial frontal cortex

A central tenet of cognitive neuroscience is that humans build an internal model of the external world and use mental simulation of the model to perform physical inferences. Decades of human experiments have shown that behaviors in many physical reasoning tasks are consistent with predictions from the mental simulation theory. However, evidence for the defining feature of mental simulation - that neural population dynamics reflect simulations of physical states in the environment - is limited. We test the mental simulation hypothesis by combining a naturalistic ball-interception task, large-scale electrophysiology in non-human primates, and recurrent neural network modeling. We find that neurons in the monkeys' dorsomedial frontal cortex (DMFC) represent task-relevant information about the ball position in a multiplexed fashion. At a population level, the activity pattern in DMFC comprises a low-dimensional neural embedding that tracks the ball both when it is visible and invisible, serving as a neural substrate for mental simulation. A systematic comparison of different classes of task-optimized RNN models with the DMFC data provides further evidence supporting the mental simulation hypothesis. Our findings provide evidence that neural dynamics in the frontal cortex are consistent with internal simulation of external states in the environment.

The Impact of Artificial Intelligence on Healthcare: A Comprehensive Review of Advancements in Diagnostics, Treatment, and Operational Efficiency

Background and Aims

Artificial Intelligence (AI) beginning to integrate in healthcare, is ushering in a transformative era, impacting diagnostics, altering personalized treatment, and significantly improving operational efficiency. The study aims to describe AI in healthcare, including important technologies like robotics, machine learning (ML), deep learning (DL), and natural language processing (NLP), and to investigate how these technologies are used in patient interaction, predictive analytics, and remote monitoring. The goal of this review is to present a thorough analysis of AI's effects on healthcare while providing stakeholders with a road map for navigating this changing environment.

Methods

This review analyzes the impact of AI on healthcare using data from the Web of Science (2014-2024), focusing on keywords like AI, ML, and healthcare applications. It examines the uses and effects of AI on healthcare by synthesizing recent literature and real-world case studies, such as Google Health and IBM Watson Health, highlighting AI technologies, their useful applications, and the difficulties in putting them into practice, including problems with data security and resource limitations. The review also discusses new developments in AI, and how they can affect society.

Results

The findings demonstrate how AI is enhancing the skills of medical professionals, enhancing diagnosis, and opening the door to more individualized treatment plans, as reflected in the steady rise of AI-related healthcare publications from 158 articles (3.54%) in 2014 to 731 articles (16.33%) by 2024. Core applications like remote monitoring and predictive analytics improve operational effectiveness and patient involvement. However, there are major obstacles to the mainstream implementation of AI in healthcare, including issues with data security and budget constraints.

Conclusion

Healthcare may be transformed by AI, but its successful use requires ethical and responsible use. To meet the changing demands of the healthcare sector and guarantee the responsible application of AI technologies, the evaluation highlights the necessity of ongoing research, instruction, and multidisciplinary cooperation. In the future, integrating AI responsibly will be essential to optimizing its advantages and reducing related dangers.

Monkeys engage in visual simulation to solve complex problems

Visual simulation-i.e., using internal reconstructions of the world to experience potential future versions of events that are not currently happening-is among the most sophisticated capacities of the human mind. But is this ability in fact uniquely human? To answer this question, we tested monkeys on a series of experiments involving the "Planko" game, which we have previously used to evoke visual simulation in human participants. We found that monkeys were able to successfully play the game using a simulation strategy, predicting the trajectory of a ball through a field of planks while demonstrating a level of accuracy and behavioral signatures comparable with those of humans. Computational analyses further revealed that the monkeys' strategy while playing Planko aligned with a recurrent neural network (RNN) that approached the task using a spontaneously learned simulation strategy. Finally, we carried out awake functional magnetic resonance imaging while monkeys played Planko. We found activity in motion-sensitive regions of the monkey brain during hypothesized simulation periods, even without any perceived visual motion cues. This neural result closely mirrors previous findings from human research, suggesting a shared mechanism of visual simulation across species. Taken together, these findings challenge traditional views of animal cognition, proposing that nonhuman primates possess a complex cognitive landscape, capable of invoking imaginative and predictive mental experiences to solve complex everyday problems.

Shaping dynamical neural computations using spatiotemporal constraints

Dynamics play a critical role in computation. The principled evolution of states over time enables both biological and artificial networks to represent and integrate information to make decisions. In the past few decades, significant multidisciplinary progress has been made in bridging the gap between how we understand biological versus artificial computation, including how insights gained from one can translate to the other. Research has revealed that neurobiology is a key determinant of brain network architecture, which gives rise to spatiotemporally constrained patterns of activity that underlie computation. Here, we discuss how neural systems use dynamics for computation, and claim that the biological constraints that shape brain networks may be leveraged to improve the implementation of artificial neural networks. To formalize this discussion, we consider a natural artificial analog of the brain that has been used extensively to model neural computation: the recurrent neural network (RNN). In both the brain and the RNN, we emphasize the common computational substrate atop which dynamics occur-the connectivity between neurons-and we explore the unique computational advantages offered by biophysical constraints such as resource efficiency, spatial embedding, and neurodevelopment.

A transient high-dimensional geometry affords stable conjunctive subspaces for efficient action selection

Flexible action selection requires cognitive control mechanisms capable of mapping the same inputs to different output actions depending on the context. From a neural state-space perspective, this requires a control representation that separates similar input neural states by context. Additionally, for action selection to be robust and time-invariant, information must be stable in time, enabling efficient readout. Here, using EEG decoding methods, we investigate how the geometry and dynamics of control representations constrain flexible action selection in the human brain. Participants performed a context-dependent action selection task. A forced response procedure probed action selection different states in neural trajectories. The result shows that before successful responses, there is a transient expansion of representational dimensionality that separated conjunctive subspaces. Further, the dynamics stabilizes in the same time window, with entry into this stable, high-dimensional state predictive of individual trial performance. These results establish the neural geometry and dynamics the human brain needs for flexible control over behavior.

The Quest for an Integrated Set of Neural Mechanisms Underlying Object Recognition in Primates

Inferences made about objects via vision, such as rapid and accurate categorization, are core to primate cognition despite the algorithmic challenge posed by varying viewpoints and scenes. Until recently, the brain mechanisms that support these capabilities were deeply mysterious. However, over the past decade, this scientific mystery has been illuminated by the discovery and development of brain-inspired, image-computable, artificial neural network (ANN) systems that rival primates in these behavioral feats. Apart from fundamentally changing the landscape of artificial intelligence, modified versions of these ANN systems are the current leading scientific hypotheses of an integrated set of mechanisms in the primate ventral visual stream that support core object recognition. What separates brain-mapped versions of these systems from prior conceptual models is that they are sensory computable, mechanistic, anatomically referenced, and testable (SMART). In this article, we review and provide perspective on the brain mechanisms addressed by the current leading SMART models. We review their empirical brain and behavioral alignment successes and failures, discuss the next frontiers for an even more accurate mechanistic understanding, and outline the likely applications.

A Transient High-dimensional Geometry Affords Stable Conjunctive Subspaces for Efficient Action Selection

Flexible action selection requires cognitive control mechanisms capable of mapping the same inputs to different output actions depending on the context. From a neural state-space perspective, this requires a control representation that separates similar input neural states by context. Additionally, for action selection to be robust and time-invariant, information must be stable in time, enabling efficient readout. Here, using EEG decoding methods, we investigate how the geometry and dynamics of control representations constrain flexible action selection in the human brain. Participants performed a context-dependent action selection task. A forced response procedure probed action selection different states in neural trajectories. The result shows that before successful responses, there is a transient expansion of representational dimensionality that separated conjunctive subspaces. Further, the dynamics stabilizes in the same time window, with entry into this stable, high-dimensional state predictive of individual trial performance. These results establish the neural geometry and dynamics the human brain needs for flexible control over behavior.

The Artificial Intelligence-Powered New Era in Pharmaceutical Research and Development: A Review

Currently, artificial intelligence (AI), machine learning (ML), and deep learning (DL) are gaining increased interest in many fields, particularly in pharmaceutical research and development, where they assist in decision-making in complex situations. Numerous research studies and advancements have demonstrated how these computational technologies are used in various pharmaceutical research and development aspects, including drug discovery, personalized medicine, drug formulation, optimization, predictions, drug interactions, pharmacokinetics/ pharmacodynamics, quality control/quality assurance, and manufacturing processes. Using advanced modeling techniques, these computational technologies can enhance efficiency and accuracy, handle complex data, and facilitate novel discoveries within minutes. Furthermore, these technologies offer several advantages over conventional statistics. They allow for pattern recognition from complex datasets, and the models, typically developed from data-driven algorithms, can predict a given outcome (model output) from a set of features (model inputs). Additionally, this review discusses emerging trends and provides perspectives on the application of AI with quality by design (QbD) and the future role of AI in this field. Ethical and regulatory considerations associated with integrating AI into pharmaceutical technology were also examined. This review aims to offer insights to researchers, professionals, and others on the current state of AI applications in pharmaceutical research and development and their potential role in the future of research and the era of pharmaceutical Industry 4.0 and 5.0.

Inferring control objectives in a virtual balancing task in humans and monkeys

Natural behaviors have redundancy, which implies that humans and animals can achieve their goals with different strategies. Given only observations of behavior, is it possible to infer the control objective that the subject is employing? This challenge is particularly acute in animal behavior because we cannot ask or instruct the subject to use a particular strategy. This study presents a three-pronged approach to infer an animal's control objective from behavior. First, both humans and monkeys performed a virtual balancing task for which different control strategies could be utilized. Under matched experimental conditions, corresponding behaviors were observed in humans and monkeys. Second, a generative model was developed that represented two main control objectives to achieve the task goal. Model simulations were used to identify aspects of behavior that could distinguish which control objective was being used. Third, these behavioral signatures allowed us to infer the control objective used by human subjects who had been instructed to use one control objective or the other. Based on this validation, we could then infer objectives from animal subjects. Being able to positively identify a subject's control objective from observed behavior can provide a powerful tool to neurophysiologists as they seek the neural mechanisms of sensorimotor coordination.

From task structures to world models: what do LLMs know?

In what sense does a large language model (LLM) have knowledge? We answer by granting LLMs 'instrumental knowledge': knowledge gained by using next-word generation as an instrument. We then ask how instrumental knowledge is related to the ordinary, 'worldly knowledge' exhibited by humans, and explore this question in terms of the degree to which instrumental knowledge can be said to incorporate the structured world models of cognitive science. We discuss ways LLMs could recover degrees of worldly knowledge and suggest that such recovery will be governed by an implicit, resource-rational tradeoff between world models and tasks. Our answer to this question extends beyond the capabilities of a particular AI system and challenges assumptions about the nature of knowledge and intelligence.

Monkeys engage in visual simulation to solve complex problems

Visual simulation - i.e., using internal reconstructions of the world to experience potential future versions of events that are not currently happening - is among the most sophisticated capacities of the human mind. But is this ability in fact uniquely human? To answer this question, we tested monkeys on a series of experiments involving the 'Planko' game, which we have previously used to evoke visual simulation in human participants. We found that monkeys were able to successfully play the game using a simulation strategy, predicting the trajectory of a ball through a field of planks while demonstrating a level of accuracy and behavioral signatures comparable to humans. Computational analyses further revealed that the monkeys' strategy while playing Planko aligned with a recurrent neural network (RNN) that approached the task using a spontaneously learned simulation strategy. Finally, we carried out awake functional magnetic resonance imaging while monkeys played Planko. We found activity in motion-sensitive regions of the monkey brain during hypothesized simulation periods, even without any perceived visual motion cues. This neural result closely mirrors previous findings from human research, suggesting a shared mechanism of visual simulation across species. In all, these findings challenge traditional views of animal cognition, proposing that nonhuman primates possess a complex cognitive landscape, capable of invoking imaginative and predictive mental experiences to solve complex everyday problems.

Shaping dynamical neural computations using spatiotemporal constraints

Dynamics play a critical role in computation. The principled evolution of states over time enables both biological and artificial networks to represent and integrate information to make decisions. In the past few decades, significant multidisciplinary progress has been made in bridging the gap between how we understand biological versus artificial computation, including how insights gained from one can translate to the other. Research has revealed that neurobiology is a key determinant of brain network architecture, which gives rise to spatiotemporally constrained patterns of activity that underlie computation. Here, we discuss how neural systems use dynamics for computation, and claim that the biological constraints that shape brain networks may be leveraged to improve the implementation of artificial neural networks. To formalize this discussion, we consider a natural artificial analog of the brain that has been used extensively to model neural computation: the recurrent neural network (RNN). In both the brain and the RNN, we emphasize the common computational substrate atop which dynamics occur-the connectivity between neurons-and we explore the unique computational advantages offered by biophysical constraints such as resource efficiency, spatial embedding, and neurodevelopment.

Inferring control objectives in a virtual balancing task in humans and monkeys

Natural behaviors have redundancy, which implies that humans and animals can achieve their goals with different control objectives. Given only observations of behavior, is it possible to infer the control strategy that the subject is employing? This challenge is particularly acute in animal behavior because we cannot ask or instruct the subject to use a particular control strategy. This study presents a threepronged approach to infer an animal's control strategy from behavior. First, both humans and monkeys performed a virtual balancing task for which different control objectives could be utilized. Under matched experimental conditions, corresponding behaviors were observed in humans and monkeys. Second, a generative model was developed that represented two main control strategies to achieve the task goal. Model simulations were used to identify aspects of behavior that could distinguish which control objective was being used. Third, these behavioral signatures allowed us to infer the control objective used by human subjects who had been instructed to use one control objective or the other. Based on this validation, we could then infer strategies from animal subjects. Being able to positively identify a subject's control objective from behavior can provide a powerful tool to neurophysiologists as they seek the neural mechanisms of sensorimotor coordination.

Reconstructing computational system dynamics from neural data with recurrent neural networks

Computational models in neuroscience usually take the form of systems of differential equations. The behaviour of such systems is the subject of dynamical systems theory. Dynamical systems theory provides a powerful mathematical toolbox for analysing neurobiological processes and has been a mainstay of computational neuroscience for decades. Recently, recurrent neural networks (RNNs) have become a popular machine learning tool for studying the non-linear dynamics of neural and behavioural processes by emulating an underlying system of differential equations. RNNs have been routinely trained on similar behavioural tasks to those used for animal subjects to generate hypotheses about the underlying computational mechanisms. By contrast, RNNs can also be trained on the measured physiological and behavioural data, thereby directly inheriting their temporal and geometrical properties. In this way they become a formal surrogate for the experimentally probed system that can be further analysed, perturbed and simulated. This powerful approach is called dynamical system reconstruction. In this Perspective, we focus on recent trends in artificial intelligence and machine learning in this exciting and rapidly expanding field, which may be less well known in neuroscience. We discuss formal prerequisites, different model architectures and training approaches for RNN-based dynamical system reconstructions, ways to evaluate and validate model performance, how to interpret trained models in a neuroscience context, and current challenges.

Neural Foundations of Mental Simulation: Future Prediction of Latent Representations on Dynamic Scenes

Humans and animals have a rich and flexible understanding of the physical world, which enables them to infer the underlying dynamical trajectories of objects and events, plausible future states, and use that to plan and anticipate the consequences of actions. However, the neural mechanisms underlying these computations are unclear. We combine a goal-driven modeling approach with dense neurophysiological data and high-throughput human behavioral readouts that contain thousands of comparisons to directly impinge on this question. Specifically, we construct and evaluate several classes of sensory-cognitive networks to predict the future state of rich, ethologically-relevant environments, ranging from self-supervised end-to-end models with pixel-wise or object-slot objectives, to models that future predict in the latent space of purely static image-pretrained or dynamic video-pretrained foundation models. We find that "scale is not all you need", and that many state-of-the-art machine learning models fail to perform well on our neural and behavioral benchmarks for future prediction. In fact, only one class of models matches these data well overall. We find that neural responses are currently best predicted by models trained to predict the future state of their environment in the latent space of pretrained foundation models optimized for dynamic scenes in a self-supervised manner. These models also approach the neurons' ability to predict the environmental state variables that are visually hidden from view, despite not being explicitly trained to do so. Finally, we find that not all foundation model latents are equal. Notably, models that future predict in the latent space of video foundation models that are optimized to support a diverse range of egocentric sensorimotor tasks, reasonably match both human behavioral error patterns and neural dynamics across all environmental scenarios that we were able to test. Overall, these findings suggest that the neural mechanisms and behaviors of primate mental simulation have strong inductive biases associated with them, and are thus far most consistent with being optimized to future predict on reusable visual representations that are useful for Embodied AI more generally.

Artificial Intelligence in Pharmaceutical Technology and Drug Delivery Design

Artificial intelligence (AI) has emerged as a powerful tool that harnesses anthropomorphic knowledge and provides expedited solutions to complex challenges. Remarkable advancements in AI technology and machine learning present a transformative opportunity in the drug discovery, formulation, and testing of pharmaceutical dosage forms. By utilizing AI algorithms that analyze extensive biological data, including genomics and proteomics, researchers can identify disease-associated targets and predict their interactions with potential drug candidates. This enables a more efficient and targeted approach to drug discovery, thereby increasing the likelihood of successful drug approvals. Furthermore, AI can contribute to reducing development costs by optimizing research and development processes. Machine learning algorithms assist in experimental design and can predict the pharmacokinetics and toxicity of drug candidates. This capability enables the prioritization and optimization of lead compounds, reducing the need for extensive and costly animal testing. Personalized medicine approaches can be facilitated through AI algorithms that analyze real-world patient data, leading to more effective treatment outcomes and improved patient adherence. This comprehensive review explores the wide-ranging applications of AI in drug discovery, drug delivery dosage form designs, process optimization, testing, and pharmacokinetics/pharmacodynamics (PK/PD) studies. This review provides an overview of various AI-based approaches utilized in pharmaceutical technology, highlighting their benefits and drawbacks. Nevertheless, the continued investment in and exploration of AI in the pharmaceutical industry offer exciting prospects for enhancing drug development processes and patient care.

Parametric control of flexible timing through low-dimensional neural manifolds

Biological brains possess an unparalleled ability to adapt behavioral responses to changing stimuli and environments. How neural processes enable this capacity is a fundamental open question. Previous works have identified two candidate mechanisms: a low-dimensional organization of neural activity and a modulation by contextual inputs. We hypothesized that combining the two might facilitate generalization and adaptation in complex tasks. We tested this hypothesis in flexible timing tasks where dynamics play a key role. Examining trained recurrent neural networks, we found that confining the dynamics to a low-dimensional subspace allowed tonic inputs to parametrically control the overall input-output transform, enabling generalization to novel inputs and adaptation to changing conditions. Reverse-engineering and theoretical analyses demonstrated that this parametric control relies on a mechanism where tonic inputs modulate the dynamics along non-linear manifolds while preserving their geometry. Comparisons with data from behaving monkeys confirmed the behavioral and neural signatures of this mechanism.