Analyzing Knowledge Retrieval Impairments Associated with Alzheimer's Disease Using Network Analyses

Visualizing and Quantifying Longitudinal Changes in Verbal Fluency Using Recurrence Plots

The verbal fluency task, where participants name as many instances of a specific semantic or phonemic category as possible in a certain time limit, is widely used to probe language and memory retrieval functions in research and clinical settings. More recently, interests in using longitudinal observations in verbal fluency to examine changes over the lifespan have grown, in part due to the increasing availability of such datasets, yet quantitative methods for comparing repeated measures of verbal fluency responses remain scarce. As a result, existing studies tend to focus only on the number of unique words produced and how this metric changes over time, overlooking changes in other important features in the data, such as the identity of the words and the order in which they are produced. Here, we provide an example of how the literature of recurrence analysis, which aims to visualize and analyze non-linear time series, may present useful visualization and analytical approaches for this problem. Drawing on this literature, we introduce a novel metric (the "distance from diagonal," or DfD) to quantify semantic fluency data that incorporates analysis of the sequence order and changes between two lists. As a demonstration, we apply these methods to a longitudinal dataset of semantic fluency in people with Alzheimer's disease and age-matched controls. We show that DfD differs significantly between healthy controls and Alzheimer's disease patients, and that it complements common existing metrics in diagnostic prediction. Our visualization method also allows incorporation of other less common metrics-including the order that words are recalled, repetitions of words within a list, and out-of-category intrusions. Additionally, we show that these plots can be used to visualize and compare aggregate recall data at the group level. These methods can improve understanding of verbal fluency deficits observed in various neuropsychiatric and neurological disorders.

Knowledge Representations Derived From Semantic Fluency Data

The semantic fluency task is commonly used as a measure of one’s ability to retrieve semantic concepts. While performance is typically scored by counting the total number of responses, the ordering of responses can be used to estimate how individuals or groups organize semantic concepts within a category. I provide an overview of this methodology, using Alzheimer’s disease as a case study for how the approach can help advance theoretical questions about the nature of semantic representation. However, many open questions surrounding the validity and reliability of this approach remain unresolved.

Neuropsychological test validation of speech markers of cognitive impairment in the Framingham Cognitive Aging Cohort

AIM

Although clinicians primarily diagnose dementia based on a combination of metrics such as medical history and formal neuropsychological tests, recent work using linguistic analysis of narrative speech to identify dementia has shown promising results. We aim to build upon research by Thomas JA & Burkardt HA et al. (J Alzheimers Dis. 2020;76:905-22) and Alhanai et al. (arXiv:1710.07551v1. 2020) on the Framingham Heart Study (FHS) Cognitive Aging Cohort by 1) demonstrating the predictive capability of linguistic analysis in differentiating cognitively normal from cognitively impaired participants and 2) comparing the performance of the original linguistic features with the performance of expanded features.

METHODS

Data were derived from a subset of the FHS Cognitive Aging Cohort. We analyzed a sub-selection of 98 participants, which provided 127 unique audio files and clinical observations (n = 127, female = 47%, cognitively impaired = 43%). We built on previous work which extracted original linguistic features from transcribed audio files by extracting expanded features. We used both feature sets to train logistic regression classifiers to distinguish cognitively normal from cognitively impaired participants and compared the predictive power of the original and expanded linguistic feature sets, and participants' Mini-Mental State Examination (MMSE) scores.

RESULTS

Based on the area under the receiver-operator characteristic curve (AUC) of the models, both the original (AUC = 0.882) and expanded (AUC = 0.883) feature sets outperformed MMSE (AUC = 0.870) in classifying cognitively impaired and cognitively normal participants. Although the original and expanded feature sets had similar AUC, the expanded feature set showed better positive and negative predictive value [expanded: positive predictive value (PPV) = 0.738, negative predictive value (NPV) = 0.889; original: PPV = 0.701, NPV = 0.869].

CONCLUSIONS

Linguistic analysis has been shown to be a potentially powerful tool for clinical use in classifying cognitive impairment. This study expands the work of several others, but further studies into the plausibility of speech analysis in clinical use are vital to ensure the validity of speech analysis for clinical classification of cognitive impairment.

What Can Network Science Tell Us About Phonology and Language Processing?

Contemporary psycholinguistic models place significant emphasis on the cognitive processes involved in the acquisition, recognition, and production of language but neglect many issues related to the representation of language‐related information in the mental lexicon. In contrast, a central tenet of network science is that the structure of a network influences the processes that operate in that system, making process and representation inextricably connected. Here, we consider how the structure found across phonological networks of several languages from different language families may influence language processing as we age and experience diseases that affect cognition during the typical and atypical acquisition of new words, during typical perception and production of speech in adults, and during language change over time. We conclude that the network science approach may not only provide insights into specific language processes but also provide a way to connect the work from these domains, which are becoming increasingly balkanized.

Methodological Considerations for Incorporating Clinical Data Into a Network Model of Retrieval Failures

Difficulty retrieving information (e.g., words) from memory is prevalent in neurogenic communication disorders (e.g., aphasia and dementia). Theoretical modeling of retrieval failures often relies on clinical data, despite methodological limitations (e.g., locus of retrieval failure, heterogeneity of individuals, and progression of disorder/disease). Techniques from network science are naturally capable of handling these limitations. This paper reviews recent work using a multiplex lexical network to account for word retrieval failures and highlights how network science can address the limitations of clinical data. Critically, any model we employ could impact clinical practice and patient lives, harkening the need for theoretically well-informed network models.

Revealing the Role of Divergent Thinking and Fluid Intelligence in Children's Semantic Memory Organization

The current theories suggest the fundamental role of semantic memory in creativity, mediating bottom-up (divergent thinking) and top-down (fluid intelligence) cognitive processes. However, the relationship between creativity, intelligence, and the organization of the semantic memory remains poorly-characterized in children. We investigated the ways in which individual differences in children's semantic memory structures are influenced by their divergent thinking and fluid intelligence abilities. The participants (mean age 10) were grouped by their levels (high/low) of divergent thinking and fluid intelligence. We applied a recently-developed Network Science approach in order to examine group-based semantic memory graphs. Networks were constructed from a semantic fluency task. The results revealed that divergent thinking abilities are related to a more flexible structure of the semantic network, while fluid intelligence corresponds to a more structured semantic network, in line with the previous findings from the adult sample. Our findings confirm the crucial role of semantic memory organization in creative performance, and demonstrate that this phenomenon can be traced back to childhood. Finally, we also corroborate the network science methodology as a valid approach to the study of creative cognition in the developmental population.

Evidence against a relation between bilingualism and creativity

Are bilinguals more creative than monolinguals? Some prior research suggests bilinguals are more creative because the knowledge representations for their second language are similarly structured to those of highly creative people. However, there is contrasting research showing that the knowledge representations of bilinguals' second language are actually structured like those of less creative people. Finally, there is growing skepticism about there being differences between bilinguals and monolinguals on non-language tasks (e.g., the bilingual advantage for executive control). We tested whether bilinguals tested in their second language are more or less creative than both monolinguals and bilinguals tested in their first language. Participants also took a repeated semantic fluency test that we used to estimate individual semantic networks for each participant. We analyzed our results with Bayesian statistics and found support for the null hypothesis that bilingualism offers no advantage for creativity. Further, using best practices for estimating semantic networks, we found support for the hypothesis that there is no association between an individual's semantic network and their creativity. This is in contrast with published research, and suggests that some of those findings may have been the result of idiosyncrasies, outdated methods for estimating semantic networks, or statistical noise. Our results call into question reported relations between bilingualism and creativity, as well as semantic network structure as an explanatory mechanism for individual differences in creativity.

Applications of Network Science to Education Research: Quantifying Knowledge and the Development of Expertise through Network Analysis

A fundamental goal of education is to inspire and instill deep, meaningful, and long-lasting conceptual change within the knowledge landscapes of students. This commentary posits that the tools of network science could be useful in helping educators achieve this goal in two ways. First, methods from cognitive psychology and network science could be helpful in quantifying and analyzing the structure of students’ knowledge of a given discipline as a knowledge network of interconnected concepts. Second, network science methods could be relevant for investigating the developmental trajectories of knowledge structures by quantifying structural change in knowledge networks, and potentially inform instructional design in order to optimize the acquisition of meaningful knowledge as the student progresses from being a novice to an expert in the subject. This commentary provides a brief introduction to common network science measures and suggests how they might be relevant for shedding light on the cognitive processes that underlie learning and retrieval, and discusses ways in which generative network growth models could inform pedagogical strategies to enable meaningful long-term conceptual change and knowledge development among students.

SNAFU: The Semantic Network and Fluency Utility

The verbal fluency task—listing words from a category or words that begin with a specific letter—is a common experimental paradigm that is used to diagnose memory impairments and to understand how we store and retrieve knowledge. Data from the verbal fluency task are analyzed in many different ways, often requiring manual coding that is time intensive and error-prone. Researchers have also used fluency data from groups or individuals to estimate semantic networks—latent representations of semantic memory that describe the relations between concepts—that further our understanding of how knowledge is encoded. However computational methods used to estimate networks are not standardized and can be difficult to implement, which has hindered widespread adoption. We present SNAFU: the Semantic Network and Fluency Utility, a tool for estimating networks from fluency data and automatizing traditional fluency analyses, including counting cluster switches and cluster sizes, intrusions, perseverations, and word frequencies. In this manuscript, we provide a primer on using the tool, illustrate its application by creating a semantic network for foods, and validate the tool by comparing results to trained human coders using multiple datasets.

Forma mentis networks quantify crucial differences in STEM perception between students and experts

In order to investigate how high school students and researchers perceive science-related (STEM) subjects, we introduce forma mentis networks. This framework models how people conceptually structure their stance, mindset or forma mentis toward a given topic. In this study, we build forma mentis networks revolving around STEM and based on psycholinguistic data, namely free associations of STEM concepts (i.e., which words are elicited first and associated by students/researchers reading "science"?) and their valence ratings concepts (i.e., is "science" perceived as positive, negative or neutral by students/researchers?). We construct separate networks for (Ns = 159) Italian high school students and (Nr = 59) interdisciplinary professionals and researchers in order to investigate how these groups differ in their conceptual knowledge and emotional perception of STEM. Our analysis of forma mentis networks at various scales indicate that, like researchers, students perceived "science" as a strongly positive entity. However, differently from researchers, students identified STEM subjects like "physics" and "mathematics" as negative and associated them with other negative STEM-related concepts. We call this surrounding of negative associations a negative emotional aura. Cross-validation with external datasets indicated that the negative emotional auras of physics, maths and statistics in the students' forma mentis network related to science anxiety. Furthermore, considering the semantic associates of "mathematics" and "physics" revealed that negative auras may originate from a bleak, dry perception of the technical methodology and mnemonic tools taught in these subjects (e.g., calculus rules). Overall, our results underline the crucial importance of emphasizing nontechnical and applied aspects of STEM disciplines, beyond purely methodological teaching. The quantitative insights achieved through forma mentis networks highlight the necessity of establishing novel pedagogic and interdisciplinary links between science, its real-world complexity, and creativity in science learning in order to enhance the impact of STEM education, learning and outreach activities.

Viability in Multiplex Lexical Networks and Machine Learning Characterizes Human Creativity

Previous studies have shown how individual differences in creativity relate to differences in the structure of semantic memory. However, the latter is only one aspect of the whole mental lexicon, a repository of conceptual knowledge that is considered to simultaneously include multiple types of conceptual similarities. In the current study, we apply a multiplex network approach to compute a representation of the mental lexicon combining semantics and phonology and examine how it relates to individual differences in creativity. This multiplex combination of 150,000 phonological and semantic associations identifies a core of words in the mental lexicon known as viable cluster, a kernel containing simpler to parse, more general, concrete words acquired early during language learning. We focus on low (N = 47) and high (N = 47) creative individuals’ performance in generating animal names during a semantic fluency task. We model this performance as the outcome of a mental navigation on the multiplex lexical network, going within, outside, and in-between the viable cluster. We find that low and high creative individuals differ substantially in their access to the viable cluster during the semantic fluency task. Higher creative individuals tend to access the viable cluster less frequently, with a lower uncertainty/entropy, reaching out to more peripheral words and covering longer multiplex network distances between concepts in comparison to lower creative individuals. We use these differences for constructing a machine learning classifier of creativity levels, which leads to an accuracy of 65 . 0 ± 0 . 9 % and an area under the curve of 68 . 0 ± 0 . 8 % , which are both higher than the random expectation of 50%. These results highlight the potential relevance of combining psycholinguistic measures with multiplex network models of the mental lexicon for modelling mental navigation and, consequently, classifying people automatically according to their creativity levels.