Skip to main content

Douglas Saddy

University of Reading, Clinical Language Sciences, Faculty Member

Followers

74

Following

30

Co-authors

3

Public Views

Supervisors: Noam Chomsky
Phone: +44 (0)118 378 6269
Address: Professor Douglas Saddy
School of Psychology and Clinical Language Sciences
CENTRE FOR INTEGRATIVE NEUROSCIENCE AND NEURODYNAMICS

UNIVERSITY OF READING
School of Psychology and Clinical Language Sciences
Earley Gate
Reading RG6 6AL

less

InterestsView All (20)

Uploads

Papers by Douglas Saddy

Subcortical restructuring as a function of multilingualism: Insights from monolinguals, bilinguals, trilinguals and quadrilinguals

Bilingualism: Language and Cognition, Dec 10, 2023

Subcortical structures implicated in language control and processing adapt structurally with incr... more Subcortical structures implicated in language control and processing adapt structurally with increasing language experience. However, the adaptation patterns across different subcortical structures remain unclear. Previous findings from bilinguals and multilinguals reveal renormalisation patterns, lending support to the Dynamic Restructuring Model (Pliatsikas, 2020). These patterns are composed of increasing volumes during the initial stages of language learning, and subsequent reductions as experience increases. T1-weighted images from 14 English monolinguals, 14 bilinguals, 14 trilinguals, and 14 quadrilinguals were obtained. The volumes of five subcortical regions implicated in language control and processing were compared amongst the groups. The findings showed group differences for every structurecaudate nucleus, nucleus accumbens, putamen, globus pallidus and thalamus. Complex patterns were unveiled for each structure, suggesting expansions and renormalisations that differ in trajectory for each group. These findings highlight the dynamic progression of subcortical adaptations, and support the notion of structural renormalisation as language experience grows. Jia'en Yee et al.

Language production and implicit statistical learning in typical development and children with acquired language disorders: an exploratory study

Statistical properties of language provide important cues for language learning and may be proces... more Statistical properties of language provide important cues for language learning and may be processed by domain-general cognitive systems. We investigated the relationship between implicit statistical learning (the unconscious detection of statistical regularities in input) and language production. Twenty typically developing (TD) children and nine children with acquired language disorders (ALD) (aged 6 to 18 years) took part in a Boston Cookie Theft picture description task. Using a computerized analysis, we investigated statistical properties, such as usage frequency of words and collocation strength of word combinations. Participants also completed a non-linguistic serial reaction time (SRT) task, which tested non-verbal, implicit statistical learning in the visual-motor modality. We determined age effects, and compared language production and SRT performance between both groups. Older TD children produced more connected language, more words, less frequent function words, more rare or novel combinations, and showed better statistical learning. Children with ALD produced less connected language, more weakly collocated combinations, displayed less lexical diversity and showed poorer statistical learning. Post-hoc analyses found correlations between statistical learning and statistical properties of spoken language. Given the rarity and heterogeneity of children with ALD, group size was small and the study should be considered exploratory. However, we note that results are compatible with the view that language production draws on statistical learning and that impairment of statistical learning can be related to language disorders.

On the Notion Maximal Projection

North East Linguistics Society, 1986

The thalamus is crucially embedded in language processing

Clinical Neurophysiology, Apr 1, 2007

body representations (hand, foot) of the right and left M1 by using a combined fMRI/ DTI-fibre tr... more body representations (hand, foot) of the right and left M1 by using a combined fMRI/ DTI-fibre tracking procedure. Methods: Using a 3 T MRI-scanner, two different body representations of M1 were identified by fMRI in five healthy volunteers (two males, three females) who were instructed to perform a Hand-Foot-paradigm in order to identify regions of interest (ROIs) for consecutive fibre tracking. FMRI, T1-MPRAGE and DTI images were fused and fibre tracking was performed after placing ROIs for the functionally defined homonymous areas for hands and feet in the right and left M1. A second ROI was placed within the area of the CC, identified by the previous tracking step and fibre tracking was performed again. Only those tracked fibres projecting into or near to the functionally defined M1 representations of both hemispheres were retained. The CC was segmented into 5 anatomical regions: rostrum/genu, anterior body, posterior body, isthmus and splenium. Results: Interhemispheric connections originating in the hand and foot area of M1 defined by fMRI were localized consistently into the posterior body and isthmus of the CC. The hand fibres overlapped with the foot fibres but were located slightly ventrally and rostrally. Conclusion: Our data demonstrate that, in humans, interhemispheric connections between defined body representations of M1 occupy circumscript parts in the posterior body and isthmus of the CC in a somatotopic fashion. This posterior shift compared to the more anterior localization of interhemispheric motor fibres in the rhesus monkey may be explained by the expansion of the human frontal cortex and its rich interhemispheric connections.

Syntax and Uncertainty

Cambridge University Press eBooks, Oct 2, 2018

Uncovering hierarchical structure through statistical dependency

In this paper, we explore the extraction of recursive nested structure in the processing of self-... more In this paper, we explore the extraction of recursive nested structure in the processing of self-similar binary sequences generated by two Lindenmayer grammars: the Fibonacci grammar and the Skip grammar. In each of these grammars only sequential order information marks the hierarchical structure. Although closely related, these grammars differ from a formal point of view: the Fibonacci grammar is perfectly scale-free and presents an isomorphism between its surface and derivational properties while the Skip grammar, although also self-similar, does not present this isomorphism. Our goal was to explore the influence of these formal differences on the extraction of hierarchical structure by the participants. To this end, we implemented these grammars in a serial reaction time task. The results show that in both the Fibonacci grammar and the Skip grammar, participants elaborated a hierarchical structure from the signal. This suggests the involvement of at least partially similar mechanisms during processing. However, some processing differences remained that cannot be explained by the hypotheses proposed so far regarding the processing of strings generated by Lsystems. We hypothesize that these effects would be due to the self-similarity of the signal which would act as a reinforcement of the structure elaborated by the participants.

Finding hierarchical structure in binary sequences : evidence from Lindenmayer grammar learning

In this article, we explore the extraction of recursive nested structure in the processing of bin... more In this article, we explore the extraction of recursive nested structure in the processing of binary sequences. Our aim was to determine whether humans learn the higher-order regularities of a highly simplified input where only sequential order information marks the hierarchical structure. To this end, we implemented sequence generated by the Fibonacci grammar in a serial reaction time task. This deterministic grammar generates aperiodic but self-similar sequences. The combination of these two properties allowed us to evaluate hierarchical learning while controlling for the use of low-level strategies like detecting recurring patterns. The deterministic aspect of the grammar allowed us to predict precisely which points in the sequence should be subject to anticipation. Results showed that participants' pattern of anticipation could not be accounted for by "flat" statistical learning processes and was consistent with them anticipating upcoming points based on hierarchical assumptions. We also found that participants were sensitive to the structure constituency, suggesting that they organized the signal into embedded constituents. We hypothesized that the participants built this structure by merging recursively deterministic transitions.

What can repetition, reading and naming tell us about Jargon aphasia?

Journal of Neurolinguistics, Feb 1, 2019

Jargon Aphasia is an acquired language disorder characterised by high proportions of nonword erro... more Jargon Aphasia is an acquired language disorder characterised by high proportions of nonword error production, rendering spoken language incomprehensible. There exist two major hypotheses relating to the source of nonword error; one implicates disruption to phonological processing and the other suggests both phonological and lexical contributions. The lexical sources are described as failure in lexical retrieval followed by surrogate phonological construction, or a lexical selection error further compounded by phonological breakdown. The current study analysed nonword error patterns of ten individuals with fluent Neologistic Jargon aphasia in word repetition, reading and picture naming to gain insights into the contributions of these different sources. It was predicted that, if lexical retrieval deficits contribute to nonword production, naming would produce a greater proportion and severity of nonword errors in comparison to repetition and reading, where phonology is present and additional sub-lexical processing can support production. Both group and case series analyses were implemented to determine whether quantity and quality of nonwords differed across the three production tasks. Nonword phoneme inventories were compared against the normative phoneme distribution to explore whether phonological production takes place within a typically organised, lexically constrained system. Results demonstrated fewer nonword errors in naming and a tendency for nonwords in naming to be characterised by lower phonological accuracy. However, nonwords were, for the most part, constructed with reference to target phonological information and, generally, nonword phonological production patterns adhered to the statistical properties of the learned phonological system. While a subset of the current group demonstrated very limited lexical processing capacity which manifested as nonword errors in naming being most disrupted, overall the results suggest that nonwords are largely underpinned by some degree of successful lexical retrieval 4 and implicate phonological sources, which manifest more severely when production is accomplished via nonlexical processing routes.

Investigations into grammatical knowledge

This work presents a case study into the nature of one brain damaged individual's ability to comp... more This work presents a case study into the nature of one brain damaged individual's ability to comprehend sentences. The focus of the research is on the ability to construe quantificational scope ambiguities. Two broad results will emerge. First, we will demonstrate that the notion of comprehension as it is currently used in aphasic research is too narrow. In addition, we will argue against the position that linguistic impairments are the result of damage to an individual's grammatical knowledge. Secondly, we will demonstrate that evidence from aphasia can bear on a formal linguistic issue in an interesting way. The aphasiological investigation (1) demonstrates that grammatical knowledge (competence) may be retained in agrammatism and (2) presents evidence of a hitherto unknown comprehension impairment that is tied to the event structure of a sentence. The linguistic analysis of the aphasialogical data argues for the existence of a syntactically active abstract argument position associated with predicates, thus providing support for an extended Davidsonian view of argument structure. This approach also provides for an alternative account of the general nature of agrammatic receptive deficits in terms of an impaired ability to distinguish the properties of such arguments and links the pattern of deficit comprehension observed to the normal range of scopal interpretations attested in languages that do not distinguish nouns and verbs at the lexical level. The particular pattern of the comprehension deficit exposed by the research leads one to conclude that universally quantified terms are understood as binding the event position in the syntactic representations generated. Definite and indefinite phrases, however, do not. The aphasic evidence suggests that indefinites are not quantificational. Never-Being a %iraduate student at MIT is a rare privilege. In the few years that I have enjoyed this status I have had the good fortune to meet many people who have had a profound impact on my life. I am grateful for their kindness, instruction, criticism, patience, impatience and good will. While working on my thesis I have benefitted from the support of an exceptional community of scholars both at MIT and elsewhere. My committee members, Noam Chomsky, Merrill Garrett, Ken Hale and Howard Lasnik have been remarkable. I am grateful to them for their guidance and support. I have also benefitted a great deal from my many meetings with Richard Larson and Jim Higginbotham. The other faculty members at MIT have played no small role in my development as a scholar,

Primary evolving networks and the comparative analysis of robust and fragile structures

Journal of Complex Networks, 2013

In this paper we consider the structure of dynamically evolving networks modelling information an... more In this paper we consider the structure of dynamically evolving networks modelling information and activity moving across a large set of vertices. We adopt the communicability concept that generalizes that of centrality which is defined for static networks. We define the primary network structure within the whole as comprising of the most influential vertices (both as senders and receivers of dynamically sequenced activity). We present a methodology based on successive vertex knockouts , up to a very small fraction of the whole primary network, that can characterize the nature of the primary network as being either relatively robust and lattice-like (with redundancies built in) or relatively fragile and tree-like (with sensitivities and few redundancies). We apply these ideas to the analysis of evolving networks derived from fMRI scans of resting human brains. We show that the estimation of performance parameters via the structure tests of the corresponding primary networks is subject to less variability than that observed across a very large population of such scans. Hence the differences within the population are significant.

Towards a classification of Lindenmayer systems

arXiv (Cornell University), Sep 27, 2018

In this paper we will attempt to classify Lindenmayer systems based on properties of sets of rule... more In this paper we will attempt to classify Lindenmayer systems based on properties of sets of rules and the kind of strings those rules generate. This classification will be referred to as a 'parametrization' of the L-space: the L-space is the phase space in which all possible L-developments are represented. This space is infinite, because there is no halting algorithm for L-grammars; but it is also subjected to hard conditions, because there are grammars and developments which are not possible states of an Lsystem: a very well-known example is the space of normal grammars. Just as the space of normal grammars is parametrized into Regular, Context-Free, Context-Sensitive, and Unrestricted (with proper containment relations holding among them; see Chomsky, 1959: Theorem 1), we contend here that the L-space is a very rich landscape of grammars which cluster into kinds that are not mutually translatable.

Processing polarity items: Contrastive licensing costs

Brain and Language, Jul 1, 2004

We describe an experiment that investigated the failure to license polarity items in German using... more We describe an experiment that investigated the failure to license polarity items in German using event-related brain potentials (ERPs). The results reveal distinct processing reflexes associated with failure to license positive polarity items in comparison to failure to license negative polarity items. Failure to license both negative and positive polarity items elicited an N400 component reflecting semantic integration cost. Failure to license positive polarity items, however, also elicited a P600 component. The additional P600 in the positive polarity violations may reflect higher processing complexity associated with a negative operator. This difference between the two types of violation suggests that the processing of negative and positive polarity items does not involve identical mechanisms.

The intractability of non-word production difficulties in jargon aphasia: Insights from therapy

Frontiers in Psychology, 2015

Enhancing dominant modes in nonstationary time series by means of the symbolic resonance analysis

Chaos, Dec 1, 2007

We present the symbolic resonance analysis ͑SRA͒ as a viable method for addressing the problem of... more We present the symbolic resonance analysis ͑SRA͒ as a viable method for addressing the problem of enhancing a weakly dominant mode in a mixture of impulse responses obtained from a nonlinear dynamical system. We demonstrate this using results from a numerical simulation with Duffing oscillators in different domains of their parameter space, and by analyzing event-related brain potentials ͑ERPs͒ from a language processing experiment in German as a representative application. In this paradigm, the averaged ERPs exhibit an N400 followed by a sentence final negativity. Contemporary sentence processing models predict a late positivity ͑P600͒ as well. We show that the SRA is able to unveil the P600 evoked by the critical stimuli as a weakly dominant mode from the covering sentence final negativity.

Language production and implicit statistical learning in typical development and children with acquired language disorders: an exploratory study

Speech, Language and Hearing, Aug 20, 2021

Statistical properties of language provide important cues for language learning and may be proces... more Statistical properties of language provide important cues for language learning and may be processed by domain-general cognitive systems. We investigated the relationship between implicit statistical learning (the unconscious detection of statistical regularities in input) and language production. Twenty typically developing (TD) children and nine children with acquired language disorders (ALD) (aged 6 to 18 years) took part in a Boston Cookie Theft picture description task. Using a computerized analysis, we investigated statistical properties, such as usage frequency of words and collocation strength of word combinations. Participants also completed a non-linguistic serial reaction time (SRT) task, which tested non-verbal, implicit statistical learning in the visual-motor modality. We determined age effects, and compared language production and SRT performance between both groups. Older TD children produced more connected language, more words, less frequent function words, more rare or novel combinations, and showed better statistical learning. Children with ALD produced less connected language, more weakly collocated combinations, displayed less lexical diversity and showed poorer statistical learning. Post-hoc analyses found correlations between statistical learning and statistical properties of spoken language. Given the rarity and heterogeneity of children with ALD, group size was small and the study should be considered exploratory. However, we note that results are compatible with the view that language production draws on statistical learning and that impairment of statistical learning can be related to language disorders.

Whoever that likes relatives

Aahus University Library eBooks, 2019

Symbolic dynamics of event-related brain potentials

Physical review, Oct 1, 2000

We apply symbolic dynamics techniques such as word statistics and measures of complexity to nonst... more We apply symbolic dynamics techniques such as word statistics and measures of complexity to nonstationary and noisy multivariate time series of electroencephalograms ͑EEG͒ in order to estimate event-related brain potentials ͑ERP͒. Their significance against surrogate data as well as between different experimental conditions is tested. These methods are validated by simulations using stochastic dynamical systems with time-dependent control parameters and compared with traditional ERP-analysis techniques. Continuous EEG data are cut into epochs according to stimuli events presented to the subjects. These ensembles of time series can be considered as ensembles of trajectories given by some dynamical systems. We employ a statistical mechanics approach motivated by the Frobenius-Perron equation and apply it to coarse-grained symbolic descriptions of the dynamics. We develop time-dependent measures of complexity founded on running cylinder sets and show that these quantities are able to distinguish simulated data obtained by different control parameters as well as experimental data between different experimental conditions. As a first finding, our approach restores the well-known ERP components and it reveals additionally qualitative changes in the EEG that cannot be detected by means of the traditional techniques. We criticize the prerequisites of the traditional approach to ERP analysis and propose to consider ERP instead in terms of dynamical system theory and information theory.

Deciphering the mechanisms of phonological therapy in jargon aphasia

International Journal of Language & Communication Disorders, Nov 26, 2018

Background: Severe word production difficulties remain one of the most challenging clinical sympt... more Background: Severe word production difficulties remain one of the most challenging clinical symptoms to treat in individuals with jargon aphasia. Clinically, it is important to determine why some individuals with jargon aphasia improve following therapy when others do not. We report a therapy study with AM, an individual with severe neologistic jargon aphasia, and provide a subsequent comparison with previous cases, with the purpose of informing both our theoretical and clinical understanding of jargon aphasia. Aims: To investigate AM's locus of word production deficit and determine the effectiveness of phonological component analysis (PCA) therapy, a phonological cueing therapy, in the re-learning and generalization of naming responses for words. In addition, AM's performance in therapy, linguistic profile and ability to engage with therapy/cues were compared in a retrospective analysis with the background linguistic and therapy data of two other individuals with jargon aphasia (P9 and FF), who responded differentially to PCA. This was undertake to explore possible prognostic indicators of phonological therapy for jargon aphasia. Methods & Procedures: A battery of linguistic and neuropsychological tests was used to identify AM's word production deficit. A single-subject multiple probe design across behaviours was employed to evaluate the effects of PCA therapy on the re-learning and generalization of naming responses. In the retrospective analysis of AM, P9 and FF, we compared differences and similarities in performance on various linguistic tasks, the ability to engage in therapy (i.e., ability to generate and use the cues), as well as to retain and maintain cues. Outcomes & Results: AM's locus of deficit was identified in the mapping between semantics and phonology. PCA was found to be effective in improving naming in two of the three treated word lists during the treatment phase; however, these gains were not maintained. Generalization to untreated picture names was not observed. Findings from the retrospective analysis illustrated that oral reading skills, the ability to segment phonological information from words and active engagement with provided cues are likely prerequisites for obtaining robust and long-term gains. Conclusions & Implications: We demonstrated that phonological therapy could be beneficial for the remediation of naming abilities at least in the re-learning phase; however, maintenance and generalization of these gains were limited. This research helps to elucidate the considerations and evaluations necessary for the appropriateness of phonological therapy and candidacy of individuals with jargon aphasia for this treatment approach.

The Human Thalamus Processes Syntactic and Semantic Language Violations

Neuron, Sep 1, 2008

Numerous linguistic operations have been assigned to cortical brain areas, but the contributions ... more Numerous linguistic operations have been assigned to cortical brain areas, but the contributions of subcortical structures to human language processing are still being discussed. Using simultaneous EEG recordings directly from deep brain structures and the scalp, we show that the human thalamus systematically reacts to syntactic and semantic parameters of auditorily presented language in a temporally interleaved manner in coordination with cortical regions. In contrast, two key structures of the basal ganglia, the globus pallidus internus and the subthalamic nucleus, were not found to be engaged in these processes. We therefore propose that syntactic and semantic language analysis is primarily realized within cortico-thalamic networks, whereas a cohesive basal ganglia network is not involved in these essential operations of language analysis.

Non-linear effects of presentation rate on sequence learning

In this article, we explore the impact of presentation rate on the extraction of hierarchical str... more In this article, we explore the impact of presentation rate on the extraction of hierarchical structure by manipulating the duration of the Response-to-Stimulus Interval (RSI) in a Serial Reaction Time (SRT) task. Multiple hypotheses have been put forward in the literature to account for the influence of RSI duration on sequence learning in the SRT task (Frensch & Miner, 1994; Huang et al., 2017; Willingham et al., 1997). However, this question has never been addressed from the perspective of hierarchical structure extraction. We found that RSI duration affected hierarchical elaboration in a non-linear way, with participants building higher hierarchical structures with an RSI of 250 ms compared to RSIs of 1000 ms and 100 ms. This finding suggests the presence of an optimal temporal window for sequence learning in the SRT task. This U-shaped effect cannot be accounted for by any of the existing hypotheses on the influence of RSI duration on sequence learning in the SRT task. We hypothesized that this effect results from the tension between the cognitive system's limited encoding capacity and the amount of information per unit of time delivered to the system.