Complex Cognitive Tasks Research Articles

Reflections on Enhancing Higher Education Classroom Effectiveness Through the Introduction of Large Language Models

Objective: The objective of this study is to explore the potential of integrating Large language models (LLMs) into higher education to enhance teaching effectiveness. It investigates how LLMs can support personalized learning, improve teacher - student interaction, and foster content innovation. The study also addresses the challenges associated with the use of LLMs, including the transformation of the teacherʼs role, data privacy concerns, and technical limitations in handling complex cognitive tasks. Methods: A mixed - method approach was used, combining a literature review, survey, and case study analysis. The literature review focused on artificial intelligence applications in education, while a survey was conducted among 120 professors and 533 students from five universities in China to gather quantitative data on their experiences with AI in education. Case studies were also analyzed to assess the effectiveness of LLM - supported learning platforms in enhancing classroom engagement, interaction, and teaching outcomes. Results: The survey results revealed that 68% of professors and 74% of students found LLMs beneficial for personalized learning and improved classroom engagement. However, 58% of professors raised concerns regarding the changing role of teachers and data privacy issues, while 49% of students worried about over - reliance on AI affecting their independent learning. Case studies showed a 30% improvement in teacher - student interaction and a 25% increase in student engagement, although LLMs struggled with advanced cognitive tasks in specialized fields such as mathematics. Conclusion: The study concludes that while LLMs offer significant advantages in improving personalized learning and enhancing interaction, their integration into higher education must be managed carefully. Teacher training, ethical considerations, and data privacy safeguards are essential. Future research should focus on optimizing LLMs for specialized academic fields and exploring their combination with emerging technologies like virtual reality and augmented reality to create more interactive learning environments.

Compositional pretraining improves computational efficiency and matches animal behavior on complex tasks.

1Recurrent neural networks (RNN) are ubiquitously used in neuroscience to capture both neural dynamics and behaviors of living systems. However, when it comes to complex cognitive tasks, training RNNs with traditional methods can prove difficult and fall short of capturing crucial aspects of animal behavior. Here we propose a principled approach for identifying and incorporating compositional tasks as part of RNN training. Taking as target a temporal wagering task previously studied in rats, we design a pretraining curriculum of simpler cognitive tasks that reflect relevant sub-computations. We show that this pretraining substantially improves learning efficacy and is critical for RNNs to adopt similar strategies as rats, including long-timescale inference of latent states, which conventional pretraining approaches fail to capture. Mechanistically, our pretraining supports the development of slow dynamical systems features needed for implementing both inference and value-based decision making. Overall, our approach is an important step for endowing RNNs with relevant inductive biases, which is important when modeling complex behaviors that rely on multiple cognitive computations.

MQGA: A quantitative analysis of brain network hubs using multi-graph theoretical indices

Recent advancements in large-scale network studies have shown that connector hubs and provincial hubs are vital for coordinating complex cognitive tasks by facilitating information transfer between and within specialized modules. However, current methods for identifying these hubs often lack standardized measurement criteria, hindering quantitative analysis. This study proposes a novel computational method utilizing multi-graph theoretical index calculations to quantitatively analyze hub attributes in brain networks. Using benchmark network, random simulation network (N = 100), resting fMRI data from the ADHD-200 NYU dataset (HC = 110, ADHD = 146), and the Peking dataset (HC = 120, ADHD = 83), we introduce the Multi-criteria Quantitative Graph Analysis (MQGA) method, which employs betweenness centrality, degree centrality, and participation coefficient to determine the connector (con) hub index and provincial (pro) hub index. The method's accuracy, reliability, and stability were validated through correlation analysis of hub indices and labels, vulnerability tests, and consistency analysis across subjects. Results indicate that as network sparsity increases, the con hub index increases while the pro hub index decreases, with the optimal hub node index at 4 % sparsity. Vulnerability tests revealed that removing con nodes had a greater impact on network integrity than removing pro nodes. Both con and pro exhibited stability in consistency analyses, but con was more stable. The stability of hub scores in disease groups was significantly lower than in the healthy control group. High con values were found in the precuneus, postcentral gyrus, and precentral gyrus, whereas high pro values were identified in the precentral gyrus, postcentral gyrus, superior parietal lobule, precuneus, and superior temporal gyrus. This approach enhances the accuracy and sensitivity of hub node identification, facilitating precise comparisons and producing consistent, replicable results, advancing our understanding of brain network hub nodes, their roles in cognitive processes, and their implications for brain disease research.

피아노 교육에서의 인지적 부하 관리 전략

This study aims to explore how piano educators perceive and manage students’ cognitive load and propose effective instructional strategies based on these findings. According to Cognitive Load Theory (CLT), learners experience cognitive load due to the limitations of working memory, and in order to improve learning outcomes, it is essential to effectively manage this cognitive load. Piano learning consists of complex cognitive tasks such as sight-reading, musical interpretation and expression, and technical skill development, which can result in a high cognitive load during the learning process. If this cognitive load is not properly managed, it can hinder learning effectiveness. In this study, focus group interviews (FGI) were conducted with college piano instructors to explore how they perceive students' cognitive load and what strategies they use to reduce it. Semi-structured interviews were employed as the research method, with the main questions focusing on the factors contributing to students' cognitive load during the learning process and the strategies instructors use to reduce cognitive load. The results revealed that instructors use various strategies to reduce students' cognitive load, such as personalized repertoire selection guidance, a step-by-step educational approach, the use of visual aids and digital tools, and promoting students' self-directed learning. The instructors found these strategies to be effective in reducing students' cognitive load during piano learning. This study holds significance by presenting various strategies that can effectively manage cognitive load in piano learning, providing practical insights that can be applied in educational settings.

Cognitive Maps for a Non-Euclidean Environment: Path Integration and Spatial Memory on a Sphere

Humans build mental models of the world and utilize them for various cognitive tasks. The exact form of cognitive maps is not fully understood, especially for novel and complex environments beyond the flat Euclidean environment. To address this gap, we investigated path integration—a critical process underlying cognitive mapping—and spatial-memory capacity on the spherical (non-Euclidean) and planar (Euclidean) environments in young healthy adults (N = 20) using immersive virtual reality. We observed a strong Euclidean bias during the path-integration task on the spherical surface, even among participants who possessed knowledge of non-Euclidean geometry. Notably, despite this bias, participants demonstrated reasonable navigation ability on the sphere. This observation and simulation suggest that humans navigate nonflat surfaces by constructing locally confined Euclidean maps and flexibly combining them. This insight sheds light on potential neural mechanisms and behavioral strategies for solving complex cognitive tasks.

Obstacle negotiation while dual-tasking in children with Developmental Coordination Disorder (DCD): An augmented-reality approach

BackgroundChildren with Developmental Coordination Disorder (DCD) exhibit deficits in predictive motor control, balance, and aspects of cognitive control, which are important for safely negotiating obstacles while walking. As concurrent performance of cognitive and motor tasks (dual-tasking) may exacerbate these deficits, we examined motor and cognitive dual-tasking differences between children with DCD and their typically developing (TD) peers during obstacle negotiation. Methods34 children aged 6–12 years (16 TD, 18 DCD) walked along a 12 m path, stepping over an obstacle (30 % or 50 % of leg length) at its mid-point. On dual-task trials, participants completed a simple or complex (cognitive) visual discrimination task presented via an augmented reality headset. Proportional dual-task costs (pDTCs) were measured on cognitive and gait outcomes over three phases: pre-obstacle, obstacle step-over, and post-obstacle. ResultsDuring the obstacle step-over phase, both groups increased their leading leg clearance when dual-tasking, while the DCD group had larger pDTC than TD for the high obstacle under simple stimulus conditions (viz simple-high combination). The complex cognitive task produced larger pDTCs than the simple one on leading leg clearance and post-obstacle gait variability. ConclusionsIn general, both DCD and TD groups showed similar pDTCs under complex conditions, while the specific deficit in DCD under the simple-high combination suggests a (default) compensatory strategy during step-over when attention is diverted to a secondary task. Competing cognitive and motor demands during obstacle negotiation present a potential safety risk for children.

Improving working memory by electrical stimulation and cross-frequency coupling

Working memory (WM) is essential for the temporary storage and processing of information required for complex cognitive tasks and relies on neuronal theta and gamma oscillations. Given the limited capacity of WM, researchers have investigated various methods to improve it, including transcranial alternating current stimulation (tACS), which modulates brain activity at specific frequencies. One particularly promising approach is theta-gamma peak-coupled-tACS (TGCp-tACS), which simulates the natural interaction between theta and gamma oscillations that occurs during cognitive control in the brain. The aim of this study was to improve WM in healthy young adults with TGCp-tACS, focusing on both behavioral and neurophysiological outcomes. Thirty-one participants completed five WM tasks under both sham and verum stimulation conditions. Electroencephalography (EEG) recordings before and after stimulation showed that TGCp-tACS increased power spectral density (PSD) in the high-gamma region at the stimulation site, while PSD decreased in the theta and delta regions throughout the cortex. From a behavioral perspective, although no significant changes were observed in most tasks, there was a significant improvement in accuracy in the 14-item Sternberg task, indicating an improvement in phonological WM. In conclusion, TGCp-tACS has the potential to promote and improve the phonological component of WM. To fully realize the cognitive benefits, further research is needed to refine the stimulation parameters and account for individual differences, such as baseline cognitive status and hormonal factors.

The neural dynamics associated with computational complexity.

Many everyday tasks require people to solve computationally complex problems. However, little is known about the effects of computational hardness on the neural processes associated with solving such problems. Here, we draw on computational complexity theory to address this issue. We performed an experiment in which participants solved several instances of the 0-1 knapsack problem, a combinatorial optimization problem, while undergoing ultra-high field (7T) functional magnetic resonance imaging (fMRI). Instances varied in computational hardness. We characterize a network of brain regions whose activation was correlated with computational complexity, including the anterior insula, dorsal anterior cingulate cortex and the intra-parietal sulcus/angular gyrus. Activation and connectivity changed dynamically as a function of complexity, in line with theoretical computational requirements. Overall, our results suggest that computational complexity theory provides a suitable framework to study the effects of computational hardness on the neural processes associated with solving complex cognitive tasks.

Implications of Developmental 17-OHPC Exposure on the Mesocorticolimbic Serotonergic and Dopaminergic Pathways and Adolescent Mood-Related Behavior in Rats.

The synthetic progestin, 17-α-hydroxyprogesterone caproate (17-OHPC), is administered to pregnant individuals at risk for recurrent preterm birth during a critical period of fetal mesocorticolimbic serotonergic and dopaminergic pathway development. These pathways play an important role in regulating cognitive behaviors later in life. Despite this, there has been very little research regarding the potential long-term effects of 17-OHPC on the behavioral and neural development of exposed children. In rodents, developmental exposure to 17-OHPC disrupts serotonergic and dopaminergic innervation of the medial prefrontal cortex and impairs decision-making in complex cognitive tasks in adulthood. The present study tested the hypothesis that developmental exposure to 17-OHPC similarly disrupts the development of serotonergic and dopaminergic pathways within limbic targets and subsequent mood-related behaviors. Developmental 17-OHPC exposure significantly increased the density of serotonin transporter-IR fibers in CA1, CA2/3, and the suprapyramidal blade of dentate gyrus in hippocampus and significantly reduced the density of TH-IR fibers within the nucleus accumbens shell in males but had no effect in females during adolescence. Irregular microglia activational phenotype and number were also observed in the hippocampus of 17-OHPC-exposed males. Developmental 17-OHPC reduced the latency to immobility in males in the forced swim test but did not affect sucrose consumption in a sucrose preference test. These findings suggest that 17-OHPC exerts sex-specific effects on the development of mesocorticolimbic pathways and mood-related behavior in adolescence and highlight the need to investigate effects in adolescent children.

High-performing teams: Is collective intelligence the answer?

The concept of a general factor of collective intelligence, proposed by Woolley et al. in 2010, has spurred interest in understanding collective intelligence within small groups. This study aims to extend this investigation by examining the validity of a general collective intelligence factor, assessing its underlying factor structure, and evaluating its utility in predicting performance on future group problem-solving tasks and academic outcomes. Employing a correlational study design, we engaged 85 university students in a series of complex cognitive tasks designed to measure collective intelligence through individual, group, and predictive phases. Contrary to the hypothesized single-factor model, our findings favor a two-factor model influenced by Cattell's theory of crystalized and fluid intelligence. These two factors accounted for substantial variance in group performance outcomes, challenging the prevailing single-factor model. Notably, the predictive validity of these factors on group assignments was statistically significant, with both individual and collective intelligence measures correlating moderately with group assignment scores (rs = .40 to .47, p < .05). Our research suggests that collective intelligence in small group settings may not be uniformly governed by a single factor but rather by multiple dimensions that reflect established theories of individual intelligence. This nuanced understanding of collective intelligence could have significant implications for enhancing group performance in both educational and organizational contexts. Future research should explore these dimensions and their independent contributions to group dynamics and outcomes.

Stylometric Analysis of AI Chatbot-Generated Emails: Are Students Losing Their Linguistic Fingerprint?

The integration of artificial intelligence (AI) in education, particularly through AI chatbots like ChatGPT, Copilot, Gemini, and Bing, has revolutionized second language teaching and learning. These chatbots, which utilize advanced natural language processing and machine learning algorithms, hold great potential. They can operate complex cognitive tasks, provide immediate feedback, enhance motivation and self-confidence, reduce anxiety and improve language skills. Despite these advantages, concerns about academic integrity, loss of unique voice, and lack of emotional depth have emerged. This research aims at conducting a stylometric analysis on 25 AI chatbot-generated emails produced by undergraduate English language learners at Sultan Qaboos University. By examining stylistic features like tone, lexical density, lexical diversity, choice of words, repetition, formality level and emotional depth, the study will provide insights into the strengths and limitations of AI chatbots in enhancing email writing skills while preserving the unique voice of student writing. Findings reveal that AI-generated emails are characterized by a repetitive structure, a high level of formality and politeness, a high lexical diversity and a lack of emotional depth and personal anecdotes. The outcomes of this research will shed light on the awareness and skills students need to acquire to optimize the use of AI chatbots without losing their linguistic fingerprint.

Eye Movements and Autonomic Regulation of Cognitive Activity during Reading in Adolescence. Part I. Functional “Cost” of Cognitive Activity when Reading Text from the Screen in Adolescence

Reading from the screen of an electronic device (ED) is a significant cognitive activity for adolescences, and its complexity affects visceral functions. We conducted an analysis of the heart rate variability (HRV) and eye movements (EM) in adolescents while they were reading complex text on an ED screen. The aim was to assess the functional state characteristics under these conditions and reveal the intensity (“cost”) associated with this activity The study involved 22 adolescences with an average age of 15 years (М = 15.46, SD = 0.44). Reading text from an ED screen in adolescents was associated with a high functional “cost,” characterized by a decrease in overall HRV, an increase in the tension index and heart rate. These changes indicate the functional tension of regulatory systems during cognitive activity. The study revealed varying levels of parafoveal processing involvement. For the majority of adolescents (86.4%), word-by-word reading and a low percentage of regressions (12.0%) were observed, suggesting developed average reading skill. However, a qualitative analysis of individual EM tracks indicated varying level of reading skill development among adolescents, possibly due to a limited vocabulary and a lack of understanding of syntax. Additionally, 40.9% of adolescents demonstrated poor text comprehension. Our results showed that reading remains a complex cognitive task for adolescents, despite the expectation that their reading skills should be well-developed and automated by this age. Individual analysis of HRV and EM in adolescents with varying levels of text comprehension during reading demonstrated different strategies of adaptive behavior and autonomic reactions when performing a complex cognitive task. The functional “cost” of information processing when reading text from the ED screen results from a combination of age-related and individual adaptation characteristics, language competence and the psycholinguistic complexity of the text.

Label-free multiphoton imaging reveals volumetric shifts across development in sensory-related brain regions of a miniature transparent vertebrate.

Animals integrate information from different sensory modalities as they mature and perform increasingly complex behaviors. This may parallel differential investment in specific brain regions depending on the demands of changing sensory inputs. To investigate developmental changes in the volume of canonical sensory integration brain regions, we used third harmonic generation imaging for morphometric analysis of forebrain and midbrain regions from 5 to 90 days post fertilization (dpf) in Danionella dracula , a transparent, miniature teleost fish whose brain is optically accessible throughout its lifespan. Relative to whole brain volume, increased volume or investment in telencephalon, a higher order sensory integration center, and torus longitudinalis (TL), a midbrain visuomotor integration center, is relatively consistent from 5 to 30 dpf, until it increases at 60 dpf, followed by another increase at 90 dpf, as animals reach adulthood. In contrast, investment in midbrain optic tectum (TeO), a retinal-recipient target, progressively decreases from 30-90 dpf, whereas investment is relatively consistent across all stages for the midbrain torus semicircularis (TS), a secondary auditory and mechanosensory lateral line center, and the olfactory bulb (OB), a direct target of the olfactory epithelium. In sum, increased investment in higher order integration centers (telencephalon, TL) occurs as juveniles reach adulthood and exhibit more complex cognitive tasks, whereas investment in modality-dominant regions occurs in earlier stages (TeO) or is relatively consistent across development (TS, OB). Complete optical access throughout Danionella 's lifespan provides a unique opportunity to investigate how changing brain structure over development correlates with changes in connectivity, microcircuitry, or behavior.

Frontoparietal Brain Network Plays a Crucial Role in Working Memory Capacity during Complex Cognitive Task.

Recent neurophysiological studies provide inconsistent results of frontoparietal network (FPN) stimulation for altering working memory (WM) capacity. This study aimed to boost WM capacity by manipulating the activity of the FPN via dual-site high-definition transcranial direct current stimulation. Forty-eight participants were randomly assigned to three stimulation groups, receiving either simultaneous anodal stimulation of the frontal and parietal areas (double stimulation), or stimulation of the frontal area only (single stimulation), or the placebo stimulation (sham) to frontal and parietal areas. After the stimulation, we used an operation span task to test memory accuracy, mathematical accuracy, time of calculation and memorizing, and recall response time across the three groups. The results revealed an enhancement of memory accuracy and a reduction of time of calculation in the double stimulation group compared with that in others. In addition, recall response time was significantly decreased in the double and single stimulation groups compared with that in sham. No differences in mathematical accuracy were observed. Our results confirm the pivotal role of the FPN in WM and suggest its functional dissociation, with the frontal component more implicated in the retrieval stage and the parietal component in the processing and retention stages.

The spontaneous activities of the multiple demand network are related to individual differences in indirect replies comprehension

This study aims to investigate the brain networks engaged in the comprehension of indirect language, as well as the individual difference in this capacity. Specially, we aim to determine whether the difference is solely influenced by the difference in individuals’ default network (DN)/language network or whether it also relies on the networks associated with processing of complex cognitive tasks, particularly the multiple demand network (MDN). Conversational indirectness scale (CIS) scores in the interpretation dimension were used as a behavioral indicator of the indirect comprehension tendency. Reading time difference between indirect replies and direct replies collected through a self-paced reading experiment was deemed as a behavioral indicator of comprehension speed of indirect replies comprehension. The two behavioral indicators were combined with resting-state functional magnetic resonance imaging (rs-fMRI). The behaviour-rfMRI analysis showed that ALFF value of right SPL and the functional connectivity (FC) between the right SPL and right IPL/SMA/ITG/Precuneus/bilateral IFG were positively correlated with the interpretation dimension of CIS scores. In addition, the ALFF value of right fusiform gyrus, the FC between the right fusiform gyrus and right precuneus, and the FCs between right SPL and right IPL/Precuneus/IFG were negatively correlated with indirect replies comprehension speed. Overlapping of these regions with large-scale brain network revealed that the right SPL was mainly located in the MDN, and the right fusiform gyrus was mainly located in the language network. Additionally, the areas showing functional connectivity with these regions were primarily located in the MDN, with a smaller subset located in the DN. Our findings suggest that the ability of individuals to actively and rapidly acquire indirect meaning relies not only on the support of the DN and the language network, but also requires collective support from the MDN.

Reconstructing human gaze behavior from EEG using inverse reinforcement learning

Decoding eye movements from non-invasive electroencephalography (EEG) data is a challenging yet vital task for both scientific and practical purposes, especially for identifying neurodegenerative disorders like Alzheimer’s disease (AD). Our research tackles this complexity by adapting inverse reinforcement learning (IRL), a machine learning method, to infer decision-making strategies from observed behaviors. We implement this to understand the processes driving eye direction and movements during diverse cognitive tasks, providing new insights into this field. Our paper begins with a detailed description of the procedures for collecting and preprocessing EEG data related to gaze behavior. We then elaborate on the development of an IRL framework designed to predict the spatial and temporal dynamics of eye movements (scanpaths) in participants engaged in cognitive tasks of varying complexity. Our model is tailored to accommodate the complexities inherent in neural signals and the stochastic nature of human gaze patterns. Our research findings underscore IRL’s effectiveness in precisely forecasting gaze patterns based on a combination of EEG and image data. The correlation between the model’s predictions and the actual gaze behavior observed in controlled experiments reinforces the utility of IRL in cognitive neuroscience research. Notably, our IRL-EEG models demonstrated superior performance, especially in more complex cognitive tasks. We further delve into the implications of our results for enhancing the understanding of neural mechanisms that govern gaze behavior.

The human subthalamic nucleus transiently inhibits active attentional processes.

The subthalamic nucleus (STN) of the basal ganglia is key to the inhibitory control of movement. Consequently, it is a primary target for the neurosurgical treatment of movement disorders like Parkinson's disease, where modulating the STN via deep brain stimulation (DBS) can release excess inhibition of thalamocortical motor circuits. However, the STN is also anatomically connected to other thalamocortical circuits, including those underlying cognitive processes like attention. Notably, STN-DBS can also affect these processes. This suggests that the STN may also contribute to the inhibition of non-motor activity and that STN-DBS may cause changes to this inhibition. Here we tested this hypothesis in humans. We used a novel, wireless outpatient method to record intracranial local field potentials (LFP) from STN DBS implants during a visual attention task (Experiment 1, n = 12). These outpatient measurements allowed the simultaneous recording of high-density EEG, which we used to derive the steady state visual evoked potential (SSVEP), a well established neural index of visual attentional engagement. By relating STN activity to this neural marker of attention (instead of overt behaviour), we avoided possible confounds resulting from STN's motor role. We aimed to test whether the STN contributes to the momentary inhibition of the SSVEP caused by unexpected, distracting sounds. Furthermore, we causally tested this association in a second experiment, where we modulated STN via DBS across two sessions of the task, spaced at least 1 week apart (n = 21, no sample overlap with Experiment 1). The LFP recordings in Experiment 1 showed that reductions of the SSVEP after distracting sounds were preceded by sound-related γ-frequency (>60 Hz) activity in the STN. Trial-to-trial modelling further showed that this STN activity statistically mediated the sounds' suppressive effect on the SSVEP. In Experiment 2, modulating STN activity via DBS significantly reduced these sound-related SSVEP reductions. This provides causal evidence for the role of the STN in the surprise-related inhibition of attention. These findings suggest that the human STN contributes to the inhibition of attention, a non-motor process. This supports a domain-general view of the inhibitory role of the STN. Furthermore, these findings also suggest a potential mechanism underlying some of the known cognitive side effects of STN-DBS treatment, especially on attentional processes. Finally, our newly established outpatient LFP recording technique facilitates the testing of the role of subcortical nuclei in complex cognitive tasks, alongside recordings from the rest of the brain, and in much shorter time than peri-surgical recordings.

Complex communicative and cognitive tasks in teaching a foreign language to students of information technology training areas

Importance. The information technology sphere has been defined by the President of the Russian Federation as one of the priorities for the development of the country’s economy and maintaining its independence status. Higher education in the field of information technology makes a significant contribution to the formation of the personnel of the industry. The purpose of education at the university is to form a highly professional personality of an IT specialist with a range of competencies. Among them, competencies related to the ability to independently obtain the necessary information, constantly update professional knowledge and experience in the rapidly developing technology sector play a special role. The need to catch up with more successful players in the field of high-tech solutions (USA, China, South Korea) determines the important role of a foreign language in the professional training of computer science students. The focus of the research is on a specific type of foreign-language educational activity – complex communicative and cognitive tasks designed to model the conditions of interrelated foreign-language communicative and intellectual activities of students.Research Methods. To conduct the research, theoretical methods are used: the study of scientific literature, analysis, generalization, synthesis of the main provisions, on the basis of which practical methods are implemented – modeling and description of students’ speech-thinking activity in the course of solving complex communicative and cognitive tasks.Results and Discussion. The theoretical justification and practical illustration of the algorithm for creating a complex communicative and cognitive task are presented. The principles of learning based on complex communicative and cognitive tasks are highlighted, an algorithm for selecting content and teaching methods are presented. A specific example of a task for the purposes of higher education is given.Conclusion. The authenticity of the quasi-professional experience obtained during the learning process can be achieved through the integration of communicative and intellectual activities in the professional field. The “brainwork-speech” dichotomy can be implemented in practical pedagogical activities on the basis of complex communicative-cognitive learning problems.

Working dogs in dynamic on-duty environments: The impact of dark adaptation, strobe lighting and acoustic distraction on task performance.

Sudden changes in sound and light (e.g., sirens and flashing police beacons) are a common component of working dogs' on-duty environment. Yet, how such stimuli impact dogs' ability to perform physical and cognitive tasks has not been explored. To address this shortcoming, we compared the accuracy and time taken for twelve dogs to complete a complex physical and cognitive task, before, during and after exposure to three 'real-world' stimuli: an acoustic distractor (85dB), white strobe lighting (5, 10 & 15 Hz), and exposure to a dazzling white, red, or blue lights. We found that strobe lighting, and to a greater extent, acoustic distraction, significantly reduced dogs' physical performance. Acoustic distraction also tended to impair dogs' cognitive performance. Dazzling lights had no effect on task performance. Most (nine out of twelve) dogs sensitised to the acoustic distraction to the extent of non-participation in the rewarded task. Our results suggest that without effective distractor response training, sudden changes in noise and flickering lights are likely to impede cognitive and physical task performance in working dogs. Repeated uncontrolled exposure may also amplify these effects.

On the ability of standard and brain-constrained deep neural networks to support cognitive superposition: a position paper

AbstractThe ability to coactivate (or “superpose”) multiple conceptual representations is a fundamental function that we constantly rely upon; this is crucial in complex cognitive tasks requiring multi-item working memory, such as mental arithmetic, abstract reasoning, and language comprehension. As such, an artificial system aspiring to implement any of these aspects of general intelligence should be able to support this operation. I argue here that standard, feed-forward deep neural networks (DNNs) are unable to implement this function, whereas an alternative, fully brain-constrained class of neural architectures spontaneously exhibits it. On the basis of novel simulations, this proof-of-concept article shows that deep, brain-like networks trained with biologically realistic Hebbian learning mechanisms display the spontaneous emergence of internal circuits (cell assemblies) having features that make them natural candidates for supporting superposition. Building on previous computational modelling results, I also argue that, and offer an explanation as to why, in contrast, modern DNNs trained with gradient descent are generally unable to co-activate their internal representations. While deep brain-constrained neural architectures spontaneously develop the ability to support superposition as a result of (1) neurophysiologically accurate learning and (2) cortically realistic between-area connections, backpropagation-trained DNNs appear to be unsuited to implement this basic cognitive operation, arguably necessary for abstract thinking and general intelligence. The implications of this observation are briefly discussed in the larger context of existing and future artificial intelligence systems and neuro-realistic computational models.