Irit Shapira-Lichter

Portraying the unique contribution of the default mode network to internally driven mnemonic processes

Numerous neuroimaging studies have implicated default mode network (DMN) involvement in both internally driven processes and memory. Nevertheless, it is unclear whether memory operations reflect a particular case of internally driven processing or alternatively involve the DMN in a distinct manner, possibly depending on memory type. This question is critical for refining neurocognitive memory theorem in the context of other endogenic processes and elucidating the functional significance of this key network. We used functional MRI to examine DMN activity and connectivity patterns while participants overtly generated words according to nonmnemonic (phonemic) or mnemonic (semantic or episodic) cues. Overall, mnemonic word fluency was found to elicit greater DMN activity and stronger within-network functional connectivity compared with nonmnemonic fluency. Furthermore, two levels of functional organization of memory retrieval were shown. First, across both mnemonic tasks, activity was greater mainly in the posterior cingulate cortex, implying selective contribution to generic aspects of memory beyond its general involvement in endogenous processes. Second, parts of the DMN showed distinct selectivity for each of the mnemonic conditions; greater recruitment of the anterior prefrontal cortex, retroesplenial cortex, and hippocampi and elevated connectivity between anterior and posterior medial DMN nodes characterized the semantic condition, whereas increased recruitment of posterior DMN components and elevated connectivity between them characterized the episodic condition. This finding emphasizes the involvement of DMN elements in discrete aspects of memory retrieval. Altogether, our results show a specific contribution of the DMN to memory processes, corresponding to the specific type of memory retrieval.

A software tool for interactive exploration of intrinsic functional connectivity opens new perspectives for brain surgery

BackgroundFunctional connectivity analysis of resting-state functional magnetic resonance imaging data (fcrs-fMRI) has been shown to be a robust non-invasive method for localization of functional networks (without using specific tasks) and to be promising for presurgical planning. However, in order to transfer the approach to everyday clinical practice, fcrs-fMRI needs to be further validated and made easily accessible to neurosurgeons. This paper addresses the latter by presenting a software tool designed for neurosurgeons for analyzing and visualizing fcrs-fMRI data.MethodsA prototypical interactive visualization tool was developed to enable neurosurgeons to explore functional connectivity data and evaluate its usability. The implementation builds upon LIPSIA, an established software package for the assessment of functional neuroimaging data, and integrates the selection of a region-of-interest with the computation and visualization of functionally connected areas. The tool was used to explore data from a healthy participant and eight brain lesion patients. The usability of the software was evaluated with four neurosurgeons previously unacquainted with the methodology, who were asked to identify prominent, large-scale cortical networks.FindingsWith this novel tool, previously published findings, such as tumor displacement of the sensorimotor cortex and other disturbances of functional networks, were reproduced. The neurosurgeons were able to consistently obtain results similar to the results of an expert, with the exception of the language network. Immediate feedback helped to pinpoint functional networks quickly and intuitively, with even inexperienced users requiring less than 3 min per network.ConclusionsAlthough fcrs-fMRI is a nascent method still undergoing evaluation with respect to established standards, the interactive software is nonetheless a promising tool for non-invasive exploration of individual functional connectivity networks in neurosurgical practice, both for well-known networks and for those less typically addressed.

Distinct functional connectivity of the hippocampus during semantic and phonemic fluency

Verbal fluency tasks are typically used in neuropsychological practice for assessment of language function in a variety of neurological disorders. Recently, it has been shown that the hippocampus, a region thought to be exclusive to the domain of memory, is also involved in tests of semantic fluency. The present study further explores hippocampal contribution to verbal fluency using functional Magnetic Resonance Imaging (fMRI) and examining mean activity and inter-regional functional connectivity with known task-related brain regions. Given the clear lateralization of brain areas involved in language, lateralization of hippocampal involvement in semantic and phonemic word fluency was also investigated. Different hippocampal recruitment during semantic and phonemic fluency was found: greater change in activity was seen during semantic fluency, as compared with phonemic fluency. This pattern was obtained in the right and the left hippocampus, with no lateralization effects. Functional connectivity analyses corroborate the notion of selective contribution of the hippocampus to semantic fluency. During the semantic fluency task, connectivity levels between the hippocampi and components of the semantic network did not differ from connectivity levels within the semantic network. In contrast, during the phonemic fluency task, the hippocampi were less correlated with components of the phonemic network, as compared to the within phonemic network connectivity. Importantly, hippocampal connectivity with the semantic network was task-dependent and restricted to periods of semantic fluency performance. Altogether, results suggest that the right and the left hippocampus are integral components of the brain network that selectively supports verbal semantic fluency, but not phonemic fluency.

Learning and memory-related brain activity dynamics are altered in systemic lupus erythematosus: a functional magnetic resonance imaging study

Background Memory impairment is prevalent in systemic lupus erythematosus (SLE); however, the pathogenesis is unknown. Methods We studied 12 patients with SLE without clinically overt neuropsychiatric manifestations and 11 matched healthy controls, aiming to characterize neural correlates of memory impairment, using structural and functional magnetic resonance imaging (MRI). The paradigm consisted of three encoding and free-recall cycles, allowing characterization of dynamics along consecutive retrieval attempts. Results During learning, patients with SLE and healthy controls showed brain activity changes in two principal networks, the default mode network (DMN) and the task-positive network (TPN). Patients with SLE demonstrated significantly less deactivation in the DMN and greater activation in the TPN, reflecting greater recruitment of both networks. The anterior medial prefrontal cortex (amPFC) of the DMN emerged as the only region where brain activity dynamics were altered both over the learning process (p < 0.006), and within free-recall period attempts (p < 0.034). Patients showed significant positive correlations between learning efficiency and hippocampal activity, and greater hippocampal functional connectivity, with pronounced connectivity to DMN structures. Conclusions Increased brain activation in patients with SLE during learning may reflect compensatory mechanisms to overcome memory impairment. Our findings localize this impairment to the amPFC, consistent with the behavioral pattern seen in SLE. Altered networking of the hippocampal subsystem of the DMN is consistent with hippocampal neuronal damage seen in SLE, and may reflect compensatory cortical reorganization to cope with dysfunction in these regions pivotal to mnemonic functions.

Inside out: a neuro-behavioral signature of free recall dynamics.

Free recall (FR) is a ubiquitous internally-driven retrieval operation that crucially affects our day-to-day life. The neural correlates of FR, however, are not sufficiently understood, partly due to the methodological challenges presented by its emerging property and endogenic nature. Using fMRI and performance measures, the neuro-behavioral correlates of FR were studied in 33 healthy participants who repeatedly encoded and retrieved word-lists. Retrieval was determined either overtly via verbal output (Experiment 1) or covertly via motor responses (Experiment 2). Brain activation during FR was characterized by two types of performance-based parametric analyses of retrieval changes over time. First was the elongation in inter response time (IRT) assumed to represent the prolongation of memory search over time, as increased effort was needed. Using a derivative of this parameter in whole brain analysis revealed the default mode network (DMN): longer IRT within FR blocks correlated with less deactivation of the DMN, representing its greater recruitment. Second was the increased number of words retrieved in repeated encoding-recall cycles, assumed to represent the learning process. Using this parameter in whole brain analysis revealed increased deactivation in the DMN (i.e., less recruitment). Together our results demonstrate the naturally occurring dynamics in the recruitment of the DMN during utilization of internally generated processes during FR. The contrasting effects of increased and decreased recruitment of the DMN following dynamics in memory search and learning, respectively, supports the idea that with learning FR is less dependent on neural operations of internally-generated processes such as those initially needed for memory search.

Impaired memory and learning abilities in patients with systemic lupus erythematosus as measured by the Rey Auditory Verbal Learning Test

Objective: The purpose of this study was to assess and characterise verbal memory impairment in patients with systemic lupus erythematosus (SLE) by the Rey Auditory Verbal Learning Test (Rey AVLT). Methods: 40 consecutive, unselected patients with SLE were evaluated with the Rey AVLT, a clinical and research tool for the study of multiple learning and memory measures. All patients were assessed for disease activity, damage, presence of antiphospholipid antibodies and depression. Findings were compared with those of 40 healthy controls matched for age, sex and education. Results: The study group included 40 patients with SLE (37 females, 3 males), median age 33 years (range 20–59), median disease duration 8 years (range 0.3–32). The median disease activity measured by the SLE Disease Activity Index (SLEDAI) was 4 (range 0–16). Median damage measured by the SLICC/ACR (Systemic Lupus International Collaborating Clinics/American College of Rheumatology) damage index score was 0 (range 0–4). Depression was detected in 16/40 patients. Several aspects of the memory domain, as measured by the Rey AVLT, were impaired in the SLE group, using analysis of variance with repeated measures. The learning curve of patients with SLE was significantly less steep compared with that of controls, (p = 0.036), the rate of words omitted from trial to trial was higher in the SLE group (p = 0.034) and retrieval was less efficient in SLE compared with controls (p = 0.004). The significance of these findings was maintained after omitting patients with stroke or depression. Conclusion: Learning ability was impaired in patients with SLE with a poor and inefficient learning strategy, as reflected by an impaired learning curve, repeated omissions and impaired retrieval. This pattern of memory deficit resembles that seen in patients with frontal lobe damage and warrants further localising brain studies.

Impaired diffusion tensor imaging findings in the corpus callosum and cingulum may underlie impaired learning and memory abilities in systemic lupus erythematosus.

Background Memory impairment is prevalent in systemic lupus erythematosus (SLE); however, its pathogenesis is unknown. In a previous functional magnetic resonance imaging (fMRI) study we demonstrated altered brain activity dynamics and less brain deactivation in patients with SLE as compared with healthy controls, when performing a learning and memory task. Our findings localized this impairment to the default mode network (DMN), and particularly to its anterior medial prefrontal cortex node. In addition, altered networking of the hippocampal subsystem of the DMN was seen in patients with SLE when performing this task, as well as atrophy of the left hippocampus. The present study aimed to search for a structural substrate for the altered recruitment pattern observed in fMRI studies using diffusion tensor imaging (DTI). Patients and methods Using DTI, we characterized brain diffusivity in 10 patients with SLE and nine healthy controls. Two tracts associated with the DMN were reconstructed: the corpus callosum (CC) and the cingulum bundle. The CC was segmented according to the Witelson segmentation scheme and the cingulum was segmented into superior and descending bundles. Results A significant increase in mean diffusivity (MD) was seen in patients with SLE without neuropsychiatric SLE (NPSLE) as compared with healthy controls in all five segments of the CC (segment 1: p = 0.043; segment 2: p = 0.005; segment 3: p = 0.003; segment 4: p = 0.012; segment 5: p = 0.023) as well as in the descending portion of the left cingulum bundle (p = 0.026). Conclusions Increased MD values in the CC and the left cingulum may indicate impaired organization/reduced integrity of these tracts, which may underlie the abnormal pattern of brain activity recruitment of the DMN observed during a verbal learning and memory task. Taking into account the central role of the left hippocampus in verbal memory, the abnormal integrity of the left cingulum may contribute to the reduced performance of patients with SLE on verbal memory tasks.

How Attention Modulates Encoding of Dynamic Stimuli

When encoding a real-life, continuous stimulus, the same neural circuits support processing and integration of prior as well as new incoming information. This ongoing interplay is modulated by attention, and is evident in regions such as the prefrontal cortex section of the task positive network (TPN), and in the posterior cingulate cortex (PCC), a hub of the default mode network (DMN). Yet the exact nature of such modulation is still unclear. To investigate this issue, we utilized an fMRI task that employed movies as the encoded stimuli and manipulated attentional load via an easy or hard secondary task that was performed simultaneously with encoding. Results showed increased intersubject correlation (inter-SC) levels when encoding movies in a condition of high, as compared to low attentional load. This was evident in bilateral ventrolateral and dorsomedial prefrontal cortices and the dorsal PCC (dPCC). These regions became more attuned to the combination of the movie and the secondary task as the attentional demand of the latter increased. Activation analyses revealed that at higher load the prefrontal TPN regions were more activated, whereas the dPCC was more deactivated. Attentional load also influenced connectivity within and between the networks. At high load the dPCC was anti-correlated to the prefrontal regions, which were more functionally coherent amongst themselves. Finally and critically, greater inter-SC in the dPCC at high load during encoding predicted lower memory strength when that information was retrieved. This association between inter-SC levels and memory strength suggest that as attentional demands increased, the dPCC was more attuned to the secondary task at the expense of the encoded stimulus, thus weakening memory for the encoded stimulus. Together, our findings show that attentional load modulated the function of core TPN and DMN regions. Furthermore, the observed relationship between memory strength and the modulation of the dPCC points to this region as a key area involved in the manipulation of attentional load on memory function.

Schema benefit vs. proactive interference: Contradicting behavioral outcomes and coexisting neural patterns

Prior knowledge can either assist or hinder the ability to learn new information. These contradicting behavioral outcomes, referred to as schema benefit and proactive interference respectively, have been studied separately. Here we examined whether the known neural correlates of each process coexist, and how they are influenced by attentional loading and aging. To this end we used an fMRI task that affected both processes simultaneously by presenting pairs of related short movies in succession. The first movie of each pair provided context for the second movie, which could evoke schema benefit and/or proactive interference. Inclusion of an easy or hard secondary task performed during encoding of the movies, as well as testing both younger (22-35y) and older (65-79y) adults, allowed examining the effect of attentional load and older age on the neural patterns associated with context. Analyses focused on three predefined regions and examined how their inter-subject correlation (inter-SC) and functional connectivity (FC) with the hippocampi changed between the first and second movie. The results in the medial prefrontal cortex (mPFC) and posterior cingulate cortex (PCC) matched and expanded previous findings: higher inter-SC and lower FC were observed during the second compared to the first movie; yet the differentiation between the first and second movies in these regions was attenuated under high attentional load, pointing to dependency on attentional resources. Instead, at high load there was a significant context effect in the FC of the left ventrolateral prefrontal cortex (vlPFC), and greater FC in the second movie was related to greater proactive interference. Further, older adults showed context effect in the PCC and vlPFC. Intriguingly, older adults with inter-SC mPFC patterns similar to younger adults exhibited schema benefit in our task, while those with inter-SC PCC patterns similar to younger adults showed proactive interference in an independent task. The brain-behavior relationships corroborate the functional significance of these regions and indicate that the mPFC mainly contributes to schema benefit, while the left vlPFC and PCC contribute to proactive interference. Importantly, our findings show that the functions of the regions are retained throughout the lifespan and may predict the predominant behavioral outcome.

Probabilistic machine learning for the evaluation of presurgical language dominance.

OBJECTIVE Providing a reliable assessment of language lateralization is an important task to be performed prior to neurosurgery in patients with epilepsy. Over the last decade, functional MRI (fMRI) has emerged as a useful noninvasive tool for language lateralization, supplementing or replacing traditional invasive methods. In standard practice, fMRI-based language lateralization is assessed qualitatively by visual inspection of fMRI maps at a specific chosen activation threshold. The purpose of this study was to develop and evaluate a new computational technique for providing the probability of each patient to be left, right, or bilateral dominant in language processing. METHODS In 76 patients with epilepsy, a language lateralization index was calculated using the verb-generation fMRI task over a wide range of activation thresholds (from a permissive threshold, analyzing all brain regions, to a harsh threshold, analyzing only the strongest activations). The data were classified using a probabilistic logistic regression method. RESULTS Concordant results between fMRI and Wada lateralization were observed in 89% of patients. Bilateral and right-dominant groups showed similar fMRI lateralization patterns differentiating them from the left-dominant group but still allowing classification in 82% of patients. CONCLUSIONS These findings present the utility of a semi-supervised probabilistic learning approach for presurgical language-dominance mapping, which may be extended to other cognitive domains such as memory and attention.