Matthew James Cooper Borkenhagen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Matthew James Cooper Borkenhagen has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Mark S. Seidenberg
Department of Psychology University of Wisconsin-Madison 53706 Madison WI
5 shared publications - Christopher Richard Cox
Department of Psychology, Louisiana State University, Baton Rouge, USA
2 shared publications Devin M. Kearns
2 shared publications- Devin M. Kearns
North Carolina State University at Raleigh
2 shared publications - Garret J. Hall
Florida State University, Tallahassee, USA
2 shared publications - Jason C. Chow
Vanderbilt University, Nashville, TN, USA
2 shared publications
Latest Publications
- Open Access
- 30 April 2026
- Published by Center for Open Science in EdArXiv
Rethinking Ground Truth on the Scoring of Oral Responses in Early Literacy Screeners
Oral-response assessments of reading and language skills have traditionally relied on human listeners to score item-level accuracy as correct or incorrect. Advances in automatic speech recognition (ASR) and speech verification systems increasingly provide item-level confidence scores bounded between 0 and 1, providing a continuous indicator of recognition/accuracy prior to enforcing any operational cut score. Our work compares a conventional two-parameter logistic (2PL) IRT model fit to human-scored dichotomies with a beta-response IRT model fit to continuous speech confidence scores from a word reading assessment. Results demonstrated that both models produced broadly comparable difficulty estimates, although the beta-response model estimated lower difficulty for a subset of items. The 2PL showed higher reliability near the average ability (θ ≈ 0) but decreased beyond θ ± 1. The beta-response model showed a more uniform measurement precision, with sustained reliability above .80 across the full ability range. These precision differences reflected how each model discriminated between ability levels. The 2PL leveraged binary accuracy to draw sharp distinctions near each item's difficulty threshold, yielding high but narrowly focused information. The beta-response model's use of continuous confidence scores allowed it to extract more information across a broader range of ability, resulting in lower peak precision but more precise measurement across the ability continuum.
- 30 April 2026
- Published by Center for Open Science in EdArXiv
Open AccessRethinking Ground Truth on the Scoring of Oral Responses in Early Literacy Screeners
Oral-response assessments of reading and language skills have traditionally relied on human listeners to score item-level accuracy as correct or incorrect. Advances in automatic speech recognition (ASR) and speech verification systems increasingly provide item-level confidence scores bounded between 0 and 1, providing a continuous indicator of recognition/accuracy prior to enforcing any operational cut score. Our work compares a conventional two-parameter logistic (2PL) IRT model fit to human-scored dichotomies with a beta-response IRT model fit to continuous speech confidence scores from a word reading assessment. Results demonstrated that both models produced broadly comparable difficulty estimates, although the beta-response model estimated lower difficulty for a subset of items. The 2PL showed higher reliability near the average ability (θ ≈ 0) but decreased beyond θ ± 1. The beta-response model showed a more uniform measurement precision, with sustained reliability above .80 across the full ability range. These precision differences reflected how each model discriminated between ability levels. The 2PL leveraged binary accuracy to draw sharp distinctions near each item's difficulty threshold, yielding high but narrowly focused information. The beta-response model's use of continuous confidence scores allowed it to extract more information across a broader range of ability, resulting in lower peak precision but more precise measurement across the ability continuum.
- Open Access
- 8 January 2026
- Published by Wiley in Reading Research Quarterly
- Vol. 61 (1)
Which Words Are Special? Identification of “Sight” Words in Educational Resources
Which words are important for early reading instruction? A standard view holds that certain words should be emphasized early in development because they are used with high frequency and/or contain atypical spelling‐sound structure. Such words have been labeled “sight,” “trick,” “snap,” or simply “high frequency” words; we refer to them as “special” words, which is intended to reflect their status as words that have been identified as being of particular instructional importance. The present study examined whether instructional resources such as commonly used curricula and word lists agree on their identity. Understanding the contents of these resources and those like them is important given their prevalence in instruction: teachers rely on wordlists to plan activities to support early word reading skills. We addressed this question using six such resources ranging from the classic Dolch list to modern commercial curricula. Results show substantial disagreement about the designated words and their properties. A total of 973 distinct words are identified in these materials. Only 28 words (3%) appear in all six resources, and over half appear in only a single one (560 words; 56%). Additional analyses demonstrate that the materials differ in terms of a number of word properties including frequency and spelling‐sound consistency. Together the results indicate a surprising lack of agreement about which words should be treated as special for instructional purposes. These differences suggest that beginning readers' learning experiences may vary greatly. In the general discussion we describe an alternative method for identifying words on a more principled basis, which would also facilitate comparisons across curricula. This method is based on computational theories of word reading that specify how properties of words such as spelling‐sound consistency and word frequency affect learning. Theories of this kind can provide a more systematic basis for identifying words to emphasize in early instruction.
- 8 January 2026
- Published by Wiley in Reading Research Quarterly
- Vol. 61 (1)
Open AccessWhich Words Are Special? Identification of “Sight” Words in Educational Resources
Which words are important for early reading instruction? A standard view holds that certain words should be emphasized early in development because they are used with high frequency and/or contain atypical spelling‐sound structure. Such words have been labeled “sight,” “trick,” “snap,” or simply “high frequency” words; we refer to them as “special” words, which is intended to reflect their status as words that have been identified as being of particular instructional importance. The present study examined whether instructional resources such as commonly used curricula and word lists agree on their identity. Understanding the contents of these resources and those like them is important given their prevalence in instruction: teachers rely on wordlists to plan activities to support early word reading skills. We addressed this question using six such resources ranging from the classic Dolch list to modern commercial curricula. Results show substantial disagreement about the designated words and their properties. A total of 973 distinct words are identified in these materials. Only 28 words (3%) appear in all six resources, and over half appear in only a single one (560 words; 56%). Additional analyses demonstrate that the materials differ in terms of a number of word properties including frequency and spelling‐sound consistency. Together the results indicate a surprising lack of agreement about which words should be treated as special for instructional purposes. These differences suggest that beginning readers' learning experiences may vary greatly. In the general discussion we describe an alternative method for identifying words on a more principled basis, which would also facilitate comparisons across curricula. This method is based on computational theories of word reading that specify how properties of words such as spelling‐sound consistency and word frequency affect learning. Theories of this kind can provide a more systematic basis for identifying words to emphasize in early instruction.
- Open Access
- 18 July 2025
- Published by Center for Open Science in PsyArXiv
The Words Matter in Early Reading Acquisition: A Simulation Study
The development of word reading skills fundamentally depends on children's interactions with printed words. A significant amount of research in the psychological and educational sciences has been conducted on the learning processes that contribute to skilled word recognition in children and adults. However, little is known about how ensembles of words encountered during early learning affect learning outcomes over timescales larger than what can be accomplished in a controlled experimental setting. In order to move past this limitation, using an established connectionist learning architecture (Plaut, McClelland, Seidenberg, & Patterson, 1996; Seidenberg & McClelland, 1989) we examined how variation in word ensembles influence learning and generalization during early reading acquisition. We simulated learners (N = 30,000) exposed to different sets of monosyllabic printed words drawn randomly from children’s literature at a large scale. In order to examine how different ensembles might impact learning depending on the learner's ability to represent the relationship between letters and sounds, we also manipulated the representational capacity of models by systematically changing the number of hidden units, holding the ensemble of words constant. Across levels of representational capacity each model had "twins" which differed only in terms of the number of hidden units they possessed. Results demonstrate significant variation in learning outcomes in a set of generalization words as a function of the word, ensemble and representational capacity. These findings inform educational practices by highlighting the importance of deliberate word selection to support early development.
- 18 July 2025
- Published by Center for Open Science in PsyArXiv
Open AccessThe Words Matter in Early Reading Acquisition: A Simulation Study
The development of word reading skills fundamentally depends on children's interactions with printed words. A significant amount of research in the psychological and educational sciences has been conducted on the learning processes that contribute to skilled word recognition in children and adults. However, little is known about how ensembles of words encountered during early learning affect learning outcomes over timescales larger than what can be accomplished in a controlled experimental setting. In order to move past this limitation, using an established connectionist learning architecture (Plaut, McClelland, Seidenberg, & Patterson, 1996; Seidenberg & McClelland, 1989) we examined how variation in word ensembles influence learning and generalization during early reading acquisition. We simulated learners (N = 30,000) exposed to different sets of monosyllabic printed words drawn randomly from children’s literature at a large scale. In order to examine how different ensembles might impact learning depending on the learner's ability to represent the relationship between letters and sounds, we also manipulated the representational capacity of models by systematically changing the number of hidden units, holding the ensemble of words constant. Across levels of representational capacity each model had "twins" which differed only in terms of the number of hidden units they possessed. Results demonstrate significant variation in learning outcomes in a set of generalization words as a function of the word, ensemble and representational capacity. These findings inform educational practices by highlighting the importance of deliberate word selection to support early development.
- Open Access
- 3 June 2025
- Published by Center for Open Science in PsyArXiv
Connectionist Models of Reading
Humans learn to read printed words by taking in perceptual information through our visual system, computing the word’s meaning, and producing a pronunciation. This process involves knowledge of language, printed and spoken, and our knowledge of the world. The connectionist formalization of this process, known as the Triangle Framework (Seidenberg & McClelland, 1989), is an influential theory of the cognitive processes associated with reading. The theory has helped explain foundational aspects of information processing, learning, typical and atypical development, and the brain bases of reading and continues to be a guiding force in reading science today.
- 3 June 2025
- Published by Center for Open Science in PsyArXiv
Open AccessConnectionist Models of Reading
Humans learn to read printed words by taking in perceptual information through our visual system, computing the word’s meaning, and producing a pronunciation. This process involves knowledge of language, printed and spoken, and our knowledge of the world. The connectionist formalization of this process, known as the Triangle Framework (Seidenberg & McClelland, 1989), is an influential theory of the cognitive processes associated with reading. The theory has helped explain foundational aspects of information processing, learning, typical and atypical development, and the brain bases of reading and continues to be a guiding force in reading science today.
- Open Access
- 1 June 2025
- Published by American Psychological Association (APA) in Canadian Journal of Experimental Psychology / Revue canadienne de psychologie expérimentale
- Vol. 79 (2) , 129-136
Parallel models of reading and numerical cognition.
Reading and math are related due to many codeveloping skills. Historically, theorizing in these two areas has progressed separately, despite well-documented empirical evidence for a range of shared underlying developmental processes subserving these learning domains. The purpose of this article was to describe the links between the Triple Code Model, an influential model of numerical cognition, and the Triangle Framework, a dominant model of learning to read. We describe several parallels between the theoretical models and discuss how the cognitive mechanisms posited by the Triangle Framework might be used to understand the commonalities in learning processes across these learning domains. In particular, we discuss how the cognitive mechanisms implemented in the Triangle Framework can be used to understand linguistic aspects of numerical cognition, specifically, learning the connections among numerals (e.g., 24) and spoken words (e.g., twenty-four), and linking those to semantic representations of magnitude. Following from these commonalities between the two models, we discuss several ways that interdisciplinary work integrating both models can benefit math cognition research. (PsycInfo Database Record (c) 2025 APA, all rights reserved).
- 1 June 2025
- Published by American Psychological Association (APA) in Canadian Journal of Experimental Psychology / Revue canadienne de psychologie expérimentale
- Vol. 79 (2) , 129-136
Open AccessParallel models of reading and numerical cognition.
Reading and math are related due to many codeveloping skills. Historically, theorizing in these two areas has progressed separately, despite well-documented empirical evidence for a range of shared underlying developmental processes subserving these learning domains. The purpose of this article was to describe the links between the Triple Code Model, an influential model of numerical cognition, and the Triangle Framework, a dominant model of learning to read. We describe several parallels between the theoretical models and discuss how the cognitive mechanisms posited by the Triangle Framework might be used to understand the commonalities in learning processes across these learning domains. In particular, we discuss how the cognitive mechanisms implemented in the Triangle Framework can be used to understand linguistic aspects of numerical cognition, specifically, learning the connections among numerals (e.g., 24) and spoken words (e.g., twenty-four), and linking those to semantic representations of magnitude. Following from these commonalities between the two models, we discuss several ways that interdisciplinary work integrating both models can benefit math cognition research. (PsycInfo Database Record (c) 2025 APA, all rights reserved).
- Open Access
- 29 January 2025
- Published by Center for Open Science in PsyArXiv
Parallel Models of Reading and Numerical Cognition
Reading and math are related due to many codeveloping skills. Historically, theorizing in these two areas has progressed separately, despite well-documented empirical evidence for a range of shared underlying developmental processes subserving these learning domains. The purpose of this paper is to describe the links between the Triple Code Model, an influential model of numerical cognition, and the Triangle Framework, a dominant model of learning to read. We describe several parallels between the theoretical models and discuss how the cognitive mechanisms posited by the Triangle Framework might be used to understand the commonalities in learning processes across these learning domains In particular, we discuss how the cognitive mechanisms implemented in the Triangle Framework can be used to understand linguistic aspects of numerical cognition, specifically, learning the connections among numerals (e.g., 24) and words (e.g., twenty-four), and linking those to semantic representations of magnitude. Following from these commonalities between the two models, we discuss several ways that interdisciplinary work integrating both models can benefit math cognition research.
- 29 January 2025
- Published by Center for Open Science in PsyArXiv
Open AccessParallel Models of Reading and Numerical Cognition
Reading and math are related due to many codeveloping skills. Historically, theorizing in these two areas has progressed separately, despite well-documented empirical evidence for a range of shared underlying developmental processes subserving these learning domains. The purpose of this paper is to describe the links between the Triple Code Model, an influential model of numerical cognition, and the Triangle Framework, a dominant model of learning to read. We describe several parallels between the theoretical models and discuss how the cognitive mechanisms posited by the Triangle Framework might be used to understand the commonalities in learning processes across these learning domains In particular, we discuss how the cognitive mechanisms implemented in the Triangle Framework can be used to understand linguistic aspects of numerical cognition, specifically, learning the connections among numerals (e.g., 24) and words (e.g., twenty-four), and linking those to semantic representations of magnitude. Following from these commonalities between the two models, we discuss several ways that interdisciplinary work integrating both models can benefit math cognition research.
- 17 June 2024
- Published by John Benjamins Publishing Company in International Journal of Chinese Linguistics
- Vol. 11 (1) , 5-29
Transitional probability between characters as a component of sentence processing in Chinese
This study explores the role of transitional probability (TP) in sentence processing in Chinese, a writing system that presents unique challenges due to its character-based structure and lack of word boundaries. The research investigates how the statistical regularities of character meaning, as captured by TP, aid in word segmentation and impact reading comprehension. Utilizing a moving window task, the study examines the processing speed of characters in high versus low TP conditions. Findings reveal that characters in high TP bigram conditions (indicating a consistent semantic association within a bigram) are processed more quickly, underscoring the importance of this statistical property of characters in Chinese sentence reading. These findings challenge conventional notions in Chinese linguistics concerning the relationship between characters, morphemes, and semantics, and suggests an alternative perspective on (and the need for reevaluation of) character-level semantics. The study also highlights the influence of prosodic context on reading speed, indicating that anticipatory linguistic patterns shape reader processing.
- 17 June 2024
- Published by John Benjamins Publishing Company in International Journal of Chinese Linguistics
- Vol. 11 (1) , 5-29
Transitional probability between characters as a component of sentence processing in Chinese
This study explores the role of transitional probability (TP) in sentence processing in Chinese, a writing system that presents unique challenges due to its character-based structure and lack of word boundaries. The research investigates how the statistical regularities of character meaning, as captured by TP, aid in word segmentation and impact reading comprehension. Utilizing a moving window task, the study examines the processing speed of characters in high versus low TP conditions. Findings reveal that characters in high TP bigram conditions (indicating a consistent semantic association within a bigram) are processed more quickly, underscoring the importance of this statistical property of characters in Chinese sentence reading. These findings challenge conventional notions in Chinese linguistics concerning the relationship between characters, morphemes, and semantics, and suggests an alternative perspective on (and the need for reevaluation of) character-level semantics. The study also highlights the influence of prosodic context on reading speed, indicating that anticipatory linguistic patterns shape reader processing.
- 5 March 2024
- Published by Wiley in The Reading Teacher
- Vol. 77 (5) , 712-726
Following the Rules in an Unruly Writing System: The Cognitive Science of Learning to Read English
- 5 March 2024
- Published by Wiley in The Reading Teacher
- Vol. 77 (5) , 712-726
Following the Rules in an Unruly Writing System: The Cognitive Science of Learning to Read English
- Open Access
Traveling waves shape neural population dynamics enabling predictions and internal model updating
The brain generates predictions based on statistical regularities in our environment. However, it is unclear how predictions are optimized through iterative interactions with the environment. Because traveling waves (TWs) propagate across the cortex shaping neural excitability, they can carry information to serve predictive processing. Using human intracranial recordings, we show that anterior-to-posterior alpha TWs correlated with prediction strength. Learning about priors altered neural state space trajectories, and how much it altered correlated with trial-by-trial prediction strength. Learning involved mismatches between predictions and sensory evidence triggering alpha-phase resets in lateral temporal cortex, accompanied by stronger alpha phase-high gamma amplitude coupling and high-gamma power. The mismatch initiated posterior-to-anterior alpha TWs and change in the subsequent trial’s state space trajectory, facilitating model updating. Our findings suggest a vital role of alpha TWs carrying both predictions to sensory cortex and mismatch signals to frontal cortex for trial-by-trial fine-tuning of predictive models.
Open AccessTraveling waves shape neural population dynamics enabling predictions and internal model updating
The brain generates predictions based on statistical regularities in our environment. However, it is unclear how predictions are optimized through iterative interactions with the environment. Because traveling waves (TWs) propagate across the cortex shaping neural excitability, they can carry information to serve predictive processing. Using human intracranial recordings, we show that anterior-to-posterior alpha TWs correlated with prediction strength. Learning about priors altered neural state space trajectories, and how much it altered correlated with trial-by-trial prediction strength. Learning involved mismatches between predictions and sensory evidence triggering alpha-phase resets in lateral temporal cortex, accompanied by stronger alpha phase-high gamma amplitude coupling and high-gamma power. The mismatch initiated posterior-to-anterior alpha TWs and change in the subsequent trial’s state space trajectory, facilitating model updating. Our findings suggest a vital role of alpha TWs carrying both predictions to sensory cortex and mismatch signals to frontal cortex for trial-by-trial fine-tuning of predictive models.
- Open Access
- 14 July 2023
- Published by Taylor & Francis in Scientific Studies of Reading
- Vol. 28 (2) , 120-141
Modeling Item-Level Spelling Variance in Adults: Providing Further Insights into Lexical Quality
PurposeThe lexical quality (LQ) hypothesis predicts that a skilled reader's lexicon will be inhabited by a range of low- to high-quality items, and the probability of representing a word with high quality varies as a function of person-level, word-level, and item-specific variables. These predictions were tested with spelling accuracy as a gauge of LQ.MethodItem-response based crossed random effects models explored simultaneous contributions of person-level (e.g. participant's decoding skill), word-level (e.g. word's transparency rating), item-specific (e.g. participant's familiarity with specific word), and person-by-word interaction predictors (e.g. decoding by transparency rating interaction) to the spelling of 25 commonly misspelled irregular English words in 61 undergraduate university students (M = 19.4 years, 70.49% female, 39.34% Hispanic, 81.97% White).ResultsSubstantial variance among individuals in item-level spelling accuracy was accounted for by person-level decoding skill; item-specific familiarity, proportion of schwas correctly represented, and correctly identifying the word from its mispronunciation; and an interaction of transparency rating by general decoding skill.ConclusionsConsistent with the LQ hypothesis, results suggest that one's ability to form a high-quality lexical representation of a given word depends on a complex combination of person-level abilities, word-level characteristics, item-specific experiences, and an interaction between person- and word-level influences.
- 14 July 2023
- Published by Taylor & Francis in Scientific Studies of Reading
- Vol. 28 (2) , 120-141
Open AccessModeling Item-Level Spelling Variance in Adults: Providing Further Insights into Lexical Quality
PurposeThe lexical quality (LQ) hypothesis predicts that a skilled reader's lexicon will be inhabited by a range of low- to high-quality items, and the probability of representing a word with high quality varies as a function of person-level, word-level, and item-specific variables. These predictions were tested with spelling accuracy as a gauge of LQ.MethodItem-response based crossed random effects models explored simultaneous contributions of person-level (e.g. participant's decoding skill), word-level (e.g. word's transparency rating), item-specific (e.g. participant's familiarity with specific word), and person-by-word interaction predictors (e.g. decoding by transparency rating interaction) to the spelling of 25 commonly misspelled irregular English words in 61 undergraduate university students (M = 19.4 years, 70.49% female, 39.34% Hispanic, 81.97% White).ResultsSubstantial variance among individuals in item-level spelling accuracy was accounted for by person-level decoding skill; item-specific familiarity, proportion of schwas correctly represented, and correctly identifying the word from its mispronunciation; and an interaction of transparency rating by general decoding skill.ConclusionsConsistent with the LQ hypothesis, results suggest that one's ability to form a high-quality lexical representation of a given word depends on a complex combination of person-level abilities, word-level characteristics, item-specific experiences, and an interaction between person- and word-level influences.