Jacqueline Cummine

FMRI of Ventral and Dorsal Processing Streams in Basic Reading Processes: Insular Sensitivity to Phonology

SummaryMost current models of the neurophysiology of basic reading processes agree on a system involving two cortical streams: a ventral stream (occipital-temporal) used when accessing familiar words encoded in lexical memory, and a dorsal stream (occipital-parietal-frontal) used when phonetically decoding words (i.e., mapping sublexical spelling onto sounds). The models diverge, however, on the issue of whether the insular cortex is involved. The present fMRI study required participants to read aloud exception words (e.g., ‘one’, which must be read via lexical memory) and pseudohomophones (e.g., ‘wun’, which must be read via sublexical spelling to sound translation) to examine the processing streams as well as the insular cortex, and their relationship to lexical and sublexical reading processes. The present study supports the notion of independent ventral-lexical and dorsal-sublexical streams, and further suggests the insular cortex to be sensitive to phonological processing (particularly sublexical spelling-sound translation). These latter findings illuminate the nature of insular activity during reading, which must be explored further in future studies, and accounted for in models of the neurophysiology of reading.

Neural representations of visual words and objects : A functional MRI study on the modularity of reading and object processing

There have been several studies supporting the notion of a ventral-dorsal distinction in the primate cortex for visual object processing, whereby the ventral stream specializes in object identification, and the dorsal stream is engaged during object localization and interaction. There is also a growing body of evidence supporting a ventral stream that specializes in lexical (i.e., whole-word) reading, and a dorsal stream that is engaged during sub-lexical reading (i.e., phonetic decoding). Here, we consider the extent to which word-reading processes are located in regions either intersecting with, or unique from, regions that sub-serve object processing along these streams. Object identification was contrasted with lexical-based reading, and object interaction processing (i.e., deciding how to interact with an object) was contrasted with sub-lexical reading. Our results suggest that object identification and lexical-based reading are largely ventral and modular, showing mainly unique regions of activation (parahippocampal and occipital-temporal gyri function associated with object identification, and lingual, lateral occipital, and posterior inferior temporal gyri function associated with lexical-based reading) and very little shared activation (posterior inferior frontal gyrus). Object interaction processing and phonetic decoding are largely dorsal, and show both modular regions of activation (more lateralized to the dorsal-frontal right hemisphere for pseudohomophone naming, and more to the dorsal-frontal left hemisphere for the object interaction task) as well as significant shared regions of processing (precentral gyri, left inferior frontal cortex, left postcentral gyrus, left lateral occipital cortex, and superior posterior temporal gyri). Given that the perceptual experimental conditions show primarily modular and very little shared processing, whereas the analytical conditions show both substantial modular and shared processing, we discuss a reconsideration of “modularity of mind” which involves a continuum between strictly modular processing and varying degrees of shared processing, and which also depends on the nature of the tasks compared (i.e., perceptual versus analytical).

Changes in Functional Magnetic Resonance Imaging Cortical Activation with Cross Education to an Immobilized Limb

PURPOSE The purpose of this study was to assess cortical activation associated with the cross-education effect to an immobilized limb, using functional magnetic resonance imaging. METHODS Fourteen right-handed participants were assigned to two groups. One group (n = 7) wore a cast and strength trained the free arm (CAST-TRAIN). The second group (n = 7) wore a cast and did not strength train (CAST). Casts were applied to the nondominant (left) wrist and hand. Strength training was maximal isometric handgrip contractions (right hand) 5 d·wk(-1). Peak force (handgrip dynamometer), muscle thickness (ultrasound), EMG, and cortical activation (functional magnetic resonance imaging) were assessed before and after the intervention. RESULTS CAST-TRAIN improved right handgrip strength by 10.7% (P < 0.01) with no change in muscle thickness. There was a significant group × time interaction for strength of the immobilized arm (P < 0.05). Handgrip strength of the immobilized arm of CAST-TRAIN was maintained, whereas the immobilized arm of CAST significantly decreased by 11% (P < 0.05). Muscle thickness of the immobilized arm decreased by an average of 3.3% (P < 0.05) for all participants and was not different between groups after adjusting for baseline differences. There was a significant group × time interaction for EMG activation (P < 0.05), where CAST-TRAIN showed an increasing trend and CAST showed a decreasing trend, pooled across arms. For the immobilized arm of CAST-TRAIN, there was a significant increase in contralateral motor cortex activation after training (P < 0.05). For the immobilized arm of CAST, there was no change in motor cortex activation. CONCLUSIONS Handgrip strength training of the free limb attenuated strength loss during unilateral immobilization. The maintenance of strength in the immobilized limb via the cross-education effect may be associated with increased motor cortex activation.

Manipulating instructions strategically affects reliance on the ventral-lexical reading stream: Converging evidence from neuroimaging and reaction time

Neurobiology of reading research has yet to explore whether reliance on the ventral-lexical stream during word reading can be enhanced by the instructed reading strategy, or whether it is impervious to such strategies. We examined Instructions: name all vs. name words (based on spelling), Word Type: regular words vs. exception words, and Word Frequency (WF) in print (log10 HAL WF) in an experiment while measuring fMRI BOLD and overt naming reaction time (RT) simultaneously. Instructions to name words increased overall reliance on the ventral-lexical stream, as measured by visible BOLD activation and the WF effect on RT, with regular words showing the greatest effects as a function of this reading strategy. Furthermore, the pattern of joint effects of these variables on RT supports the notion of cascaded, not parallel, processing. These results can be accommodated by dual-stream cascaded models of reading, and present a challenge to single-mechanism parallel processing models.

Localisation of function for noun and verb reading: Converging evidence for shared processing from fMRI activation and reaction time

Some researchers have argued in favour of verbs primarily activating the left frontal operculum (FO) in the dorsal stream, and left middle temporal (MT) region in the ventral stream, and that nouns primarily activate the left inferior temporal (IT) region in the ventral stream. Others have suggested that the activation representing noun and verb processing involves a shared neural network. We explored these hypotheses through the naming of identical, homonymous, separately cued nouns (the bat) and verbs (to bat) presented in word format using a modified naming task that ensured participants were treating the target as the appropriate part of speech (POS). Using homonymous homographs for both the noun and verb referents provides for an optimally controlled comparison given the target stimuli and responses are physically identical. Experiment 1 was a functional magnetic resonance imaging (fMRI) experiment that showed that the majority of activation was shared by both the noun and verb naming conditions, across both the ventral and dorsal streams, including the regions suggested by previous researchers as unique to verbs (FO, MT) or nouns (IT). In contrast, there was little unique activation attributable to noun processing, and practically no unique activation attributable to verb processing. This experiment, which supports a spatially shared ventral and dorsal network for noun and verb naming, was the impetus for new hypotheses involving the sharing of processes in time. Experiment 2 showed an overadditive interaction on naming reaction time (RT) between POS and bigram frequency, which provided converging evidence that the shared processing for nouns and verbs involves sublexical processing, whereas an overadditive interaction between POS and word frequency provided converging evidence that the shared processing also involves orthographic lexical access. As such, our study provides converging fMRI and RT evidence that noun and verb reading predominantly share processing along both the ventral-lexical and dorsal-sublexical reading streams.

A neuroanatomical examination of embodied cognition: semantic generation to action-related stimuli.

The theory of embodied cognition postulates that the brain represents semantic knowledge as a function of the interaction between the body and the environment. The goal of our research was to provide a neuroanatomical examination of embodied cognition using action-related pictures and words. We used functional magnetic resonance imaging (fMRI) to examine whether there were shared and/or unique regions of activation between an ecologically valid semantic generation task and a motor task in the parietal-frontocentral network (PFN), as a function of stimulus format (pictures versus words) for two stimulus types (hand and foot). Unlike other methods for neuroimaging analyses involving subtractive logic or conjoint analyses, this method first isolates shared and unique regions of activation within-participants before generating an averaged map. The results demonstrated shared activation between the semantic generation and motor tasks, which was organized somatotopically in the PFN, as well as unique activation for the semantic generation tasks in proximity to the hand or foot motor cortex. We also found unique and shared regions of activation in the PFN as a function of stimulus format (pictures versus words). These results further elucidate embodied cognition in that they show that brain regions activated during actual motor movements were also activated when an individual verbally generates action-related semantic information. Disembodied cognition theories and limitations are also discussed.

A Functional Neuroimaging Case Study of Meares–Irlen Syndrome/Visual Stress (MISViS)

The aim of this study is to evaluate the data from a participant in a reading study who had a diagnosis of Meares–Irlen syndrome/visual stress (MISViS). MISViS is characterised by visual distortions and somatic issues, which are remediated using coloured filters. The authors present a case study providing descriptive neurobiological comparisons of MISViS versus a control group. The study involved eleven English language speakers who participated in behavioural and neuroimaging versions of a language experiment with varied proportions of regular and exception words. Behavioural measures included accuracy and response times. Neuroimaging was conducted using a 1.5T Siemens Sonata MRI. The MISViS participant’s data were removed from the overall experiment and analysed as a case study. Impulse response functions (IRFs) and percentage of active voxels were extracted from four regions of interest: BAs 17, 18, 19, and the postcentral gyrus (PG) and two control regions (BA6 and left BA45). The results indicated that significant differences existed between the control group and the MISViS participant for IRF intensity in two regions (BA6 and PG) and percentage of active voxels in four regions (BA17, BA19, PG, and BA6). No significant differences occurred in left BA45 for either variable of interest. No significant differences were found for behavioural measures. In conclusion, our findings offer one of the first neurobiological descriptions of differences in IRF intensity and percentage of active voxels in visual and somatosensory cortex during a language experiment for a participant with MISViS in the absence of migraine compared to controls.

Basic reading skills and dyslexia: three decades following right versus left hemispherectomy for childhood-onset intractable epilepsy.

Dyslexia was explored within the framework of three explanations for language functioning following hemispherectomy (i.e., equipotentiality, hemispheric specialization, and crowding hypothesis/hierarchy of specialized functions) and the extent to which these models explain reading performance in S.M. (age 48, right hemispherectomy) and J.H. (age 49, left hemispherectomy). Basic reading performance was evaluated by assessing whole-word and subword reading. Both participants displayed severely impaired reading performance on pseudohomophones (e.g., WUN), signifying poor subword reading. However, J.H. (remaining right hemisphere) also demonstrated impairments in reading exception words (e.g., ONE), suggestive of poor whole-word reading. Thus, although S.M. clearly demonstrated phonological dyslexia and retention of the priority whole-word reading skills, J.H. presented with deficits more characteristic of mixed dyslexia. Taken as a whole, we suggest that some modification of the hierarchy of specialized functioning model and crowding hypothesis is needed, including stipulations about hemispheric specialization, to more accurately accommodate the present data.

An examination of the rapid automatized naming–reading relationship using functional magnetic resonance imaging

Rapid automatized naming (RAN) has been established to be a strong predictor of reading. Yet, the neural correlates underlying the RAN-reading relationship remain unknown. Thus, the purpose of this study was to determine: (a) the extent to which RAN and reading activate similar brain regions (within subjects), (b) whether RAN and reading are directly related in the shared activity network outlined in (a), and (c) to what extent RAN neural activation predicts behavioral reading performance. Using functional magnetic resonance imaging (fMRI), university students (N=15; Mean age=20.6 years) were assessed on RAN (letters and digits) and single-word reading (words and non-words). The results revealed a common RAN-reading network that included regions associated with motor planning (cerebellum), semantic access (middle temporal gyrus), articulation (supplementary motor area, pre-motor), and grapheme-phoneme translation (supramarginal gyrus). We found differences between RAN and reading with respect to percent signal change (PSC) in phonological and orthographic regions, but not in articulatory regions. Significant correlations between the neural RAN and reading parameters were found primarily in motor/articulatory regions. Further, we found a unique relationship between in-scanner reading response time and RAN PSC in the left inferior frontal gyrus. Taken together, these findings support the notion that RAN and reading activate similar neural networks. However, the relationship between RAN and reading is primarily driven by commonalities in the motor-sequencing/articulatory processes.

Localizing the Frequency × Regularity word reading interaction in the cerebral cortex

The aim of this work is to combine behavioural and functional magnetic resonance imaging (fMRI) data to advance our knowledge of where the Frequency x Regularity interaction on word naming is located in the cerebral cortex. Participants named high and low frequency, regular and exception words in a behavioural lab (Experiment 1) and during an fMRI study (Experiment 2). We used the Additive Factors Method (AFM) to localize the expected overadditive Frequency x Regularity interaction both temporally, through word naming reaction times (whereby low frequency exceptions produce the longest reaction times), and spatially on the cortex, through hemodynamic response measures from fMRI (whereby low frequency exceptions produce the highest activation intensities). Activation maps revealed significant activation for low frequency exception words in the supplementary motor association cortex (SMA). We interpret the SMA activation as increased articulatory preparation, given previous demonstrations of the SMAs involvement in motor programming. Hemodynamic time courses were extracted from four regions of interest: the middle temporal gyri, SMA, insula and the inferior frontal gyri. Importantly, hemodynamic intensities within the SMA displayed an overadditive interaction pattern parallel to that found with naming reaction times. Thus, we provide an application of the AFM to fMRI intensity measures and evidence that the SMA is a potential cortical source of the Frequency x Regularity interaction during a basic naming paradigm. While the AFM has traditionally been used to localize factors in time we provide evidence that the AFM is useful in understanding how variables influence one another in the brain.