Presentation Sentences were presented in an event-related design. The acquisition session for each participant comprised four runs of 12 min 40 s each. Each run consisted of 67 randomized trials: 51 experimental, nine catch and seven null trials. Experimental trials began with the auditory presentation of one sentence, followed by a ms interval, after which a visual fixation cross appeared for ms.
In catch trials, the fixation cross was replaced by a question mark, followed by a written sentence presented for ms, and participants were instructed to blink their eyelids once in case the written sentence matched the auditory sentence, or twice in case of mismatch for full details, see [ 24 ], where the same procedure was applied.
In order to prevent scanner noise from affecting auditory sentence presentation, fMRI sparse sampling was employed [ 25 , 26 ]. Each functional image comprised 35 axial slices 3. Each participant underwent four fMRI scanning sessions, each comprising 71 scans, plus two initial dummy scans, which were discarded prior to data analysis.
Data were preprocessed with SPM8 www. Smoothing was not performed to provide optimal sensitivity for high-frequency multi-voxel patterns in MVPA [ 27 ]. The time series of each subject were high-pass filtered at s. No pre-whitening and no global normalization was applied. Additional regressors modelled the catch trials and movement parameters. PyMVPA 2. Subject-wise z -scoring normalization was applied to correct for noise-related inhomogeneities in voxel intensities. The t -contrast images were averaged subject-wise and condition-wise. Classifications were performed between subjects in order to examine whether brain activation patterns were consistent across subjects, by means of a leave-one-subject-out cross-validation [ 24 ].
We report the mean cross-individual classification accuracies across all inclusive mask voxels i. In addition, we used searchlight analysis [ 30 ] with 4 mm radius spheres and a Gaussian Naive Bayes classifier [ 31 ] to localize anatomically the brain regions that significantly contributed to accurate discrimination of the different classification problems, as determined through a Monte Carlo permutation testing procedure. We report the mean classification accuracies across leave-one-subject-out cross-validations and the corresponding confusion matrices for the significant searchlight spheres.
For CP5, in order to gain a deeper insight into the brain coding of semantic information for the target conceptual categories, we adapted the procedure described in [ 24 ], which is based on recursive feature elimination and the sensitivity weights it provides.
Sensitivity weights reflect the contribution of each voxel to the discrimination of one category from the others [ 33 ]. For each category, we calculated the spatial intersection of all pairwise sensitivity maps involving that category e.
Log in to Wiley Online Library
Before calculation of the category-specific intersections, the pairwise sensitivity maps were filtered for a minimum cluster extension of 20 voxels, only the clusters with sensitivity weights in the 9. The four concreteness by polarity factorial combinations were classified with a mean accuracy of Figure 1.
Cell numbers represent the mean number of subjects that were classified either correctly diagonal or incorrectly off the diagonal , with respect to the four concreteness by polarity factorial combinations AA, affirmative abstract; NA, negative abstract; AC, affirmative concrete; NC, negative concrete. Colour codes are indicated by the colour palette inset. The effects are displayed on lateral and medial wall surface renderings of the average anatomical image of all participants.
Left and right hemispheres are displayed according to the neurological convention. For abstract concepts, affirmative and negative sentences were discriminated with a mean classification accuracy of Classification accuracy in individual searchlights was more successful. For concrete concepts, affirmative and negative sentences were discriminated with a mean classification accuracy of Again, classification accuracy in individual searchlights was more successful.
- Endogenous and exogenous modulators of potentials evoked by a painful cutaneous laser (LEPs);
- 1. Introduction;
- Save Me A Place In Heaven.
- Starting Your Best Life Now: A Guide for New Adventures and Stages on Your Journey (Faithwords).
- Navigation menu.
The abstract and concrete conditions partially differed with respect to the macro-anatomical distribution of the searchlights, significantly discriminating between affirmative and negative sentences electronic supplementary material, figure S1. For the abstract conditions, there was a unique involvement of the left temporal pole, right medial temporal lobe amygdala, hippocampus and parahippocampal gyrus , right precuneus, and cerebellum electronic supplementary material, table S2B. Only the concrete conditions, in turn, involved the left posterior middle temporal gyrus, the left angular gyrus, the pars opercularis of the right inferior frontal gyrus, the right superior frontal gyrus, the calcarine and lingual gyri, and, bilaterally, the postcentral gyrus and the putamen electronic supplementary material, table S2C.
The mean whole-brain accuracy for the classification of the 12 classes of sentences was However, the confusion matrix showed a meaningless structure, that is, an inconsistently populated leading diagonal, and a disproportionally high rate of densely populated off-the-diagonal cells, representing incorrect predictions-to-target correspondences electronic supplementary material, figure S2.
Therefore, no further analysis of CP2 was carried out. Affirmative and negative sentences were discriminated with a mean classification accuracy of Abstract and concrete sentences were discriminated with a mean classification accuracy of The mean whole-brain accuracy for the classification of the fine-grained conceptual categories was Searchlight permutation testing yielded a much lower mean classification accuracy Figure 2. MVPA classification of the main effect of fine-grained conceptual category. The category-specific intersections are displayed on axial slices z -coordinate levels indicated in mm of the average anatomical image of all participants neurological convention.
To further investigate this broadly distributed category-specificity we applied an alternative whole-brain approach based on sensitivity weights, yielded by recursive feature elimination. We calculated all pairwise classifications among the six categories. This study investigated how the neural processing of abstract and concrete concepts expressed at the sentence level is modulated by negation polarity.
Sentential negation polarity is thought to operate at the syntax-semantic interface [ 35 , 36 ], thus representing a case of interplay between a linguistic contextual operator and conceptual representations. Our main intent was to provide a proof of concept for the hypothesis that the neural networks supporting semantic representations are flexibly modulated by the linguistic sentential context [ 18 ]. We expanded on two different lines of research: a first line indicating that abstract and concrete concepts, and their respective fine-grained sub-categories, are distinctively encoded in distributed brain networks including category-invariant and category-specific nodes [ 3 , 4 ]; and a second line, so far limited to concrete action-related concepts, indicating that sentential negation modulates neural activity of category-specific conceptual representation nodes [ 14 — 17 ].
Post hoc classifications showed that affirmative and negative sentences were discriminable also when abstract and concrete conditions were analysed separately. Lack of sentential negation modulation on fine-grained semantic categories might be due to methodological aspects. In fact, analyses related to CP2 required separately modelling each of the 12 experimental conditions. This was not the case for CP1, where data were averaged across multiple experimental conditions. It is possible that the number of trials for each experimental condition in our study was not sufficient to ensure fully-fledged category by polarity MVPA separation in CP2.
- Prefrontal cortex - Wikipedia.
- Chapter 12 - Limbic System: Cingulate Cortex;
- Chapter 12 - Limbic System: Cingulate Cortex!
- Recent Publications | BMS | CityU.
However, this result might also constitute true negative evidence, weakening the conclusions drawn in the present study, and this important limitation needs to be considered. To sum up, the expectations of our proof of concept on negative versus affirmative linguistic contexts were fully met at a level of a simple but conceptually relevant distinction between concrete and abstract sentences, suggesting that the effect of sentential negation is not only limited to concrete action-related concepts but also extends to abstract ones.
Furthermore, successful classifications were observed when considering the main effects nested in our factorial manipulation.
Concerning sentential negation polarity, the MVPA yielded accurate whole-brain classification of the neural patterns associated with affirmative versus negative sentences although classification accuracy in individual searchlights was more successful CP3. More robust classification accuracy was observed for concreteness CP4 , and fine-grained conceptual category CP5 main effects.
Anatomical localization of condition-specific fMRI activation patterns is not straightforward in MVPA, since decoding is blind to the spatial organization of these patterns [ 37 ]. Nevertheless, MVPA localization techniques such as searchlight analysis [ 30 ], in combination with independent meta-analytic evidence on the brain functional organization of semantic processing, can provide useful information on the brain regions where contextual sentential negation modulations may occur.
The classification problems CP1 and CP3, which both addressed the manipulation of syntactic polarity, showed overlap in the pars triangularis of the inferior frontal gyrus, the basal ganglia notably, the left caudate nucleus , and the anterior and middle cingulate cortex.
Both the pars triangularis of the inferior frontal gyrus [ 38 , 39 ] and the left caudate nucleus [ 40 , 41 ] have been consistently implicated in syntactic structure processing, that is, word order computation at the sentence level above and beyond the specific issue of sentential negation. Of even greater relevance are previous univariate fMRI studies specifically investigating negation at the syntactic level, independently of the meanings on which it operates, which found an involvement of the basal ganglia [ 14 ] and of the left pars triangularis [ 22 ], among a set of other brain regions not identified in the present study.
However, the inclusion of the anterior and middle cingulate cortex as a region of overlap between CP1 and CP3 prompts another possible functional interpretation. The left caudate nucleus, in turn, is known to be crucially involved in language monitoring and control [ 43 ].
Publications Derived From HBTRC Tissue Samples (2000-2013)
It is therefore possible that these three brain regions jointly contribute to a cognitive control system that may help regulating the contextually-driven modulatory effects of negation on conceptual representations. The syntactic and cognitive control interpretations need not be mutually exclusive: the neural circuit activations underlying syntactic and cognitive control functions may be spatially and temporally distinct, but display the observed spatio-temporal overlap due to the relatively low spatial and temporal resolution of the fMRI data, and to the multivariate analysis technique.
Our results may therefore indicate that both syntactic computation and cognitive control are involved in processing negation polarity at the sentence level. In turn, the classification problems CP1 and CP4, which both addressed the experimental manipulation of semantic concreteness, showed overlap in an extended set of regions, which is more consistent with a semantic functional role. This set included the bilateral anterior temporal lobe, which has been suggested to serve as the main brain's semantic hub [ 3 , 44 ].
Spec Care Dentist , J Am Dent Assoc , Am J Psychiatry , Science , In Cochrane Library , ed 2, Oxford, Urol Res , Neuropsychopharmacology , Lancet , J Public Health Dent , J Psychiatr Neurosci , Sleep , Curr Opin Pharmacol , Accessed July 27, Biol Psychiatry , Cereb Cortex , Psychopharmacology Berl , Within this network, activity in the right precuneus reflected more detailed representations of subjective contents during vivid relative to non-vivid, recollection.
Our results suggest a more specific mechanism underlying the phenomenology of vivid autobiographical reminiscence, supported by rich subjective content representations in the precuneus, a hub of the DMN previously implicated in metacognitive evaluations during memory retrieval. Tulving 1 , 2 suggested that episodic memory is a unique human capability that enables us to engage in mental-time travel along a subjective timeline to reinstate past experiences.
In a previous study, we identified the neural correlates of the objective spatiotemporal axes along which mental travel occurs during autobiographical memory retrieval 3.
Manual Chapter , Cingulate Cortex
However, the concept of episodic memory is incomplete without a notion of the self, the accompanying subjective dimensions of experience, and a special inwardly turned state of consciousness—termed autonoetic awareness—that guides retrieval and monitoring of autobiographical memories. In this paper, we describe the networks involved in representing subjective , self-relevant content of real-world events during autobiographical reminiscence. Autobiographical memory concerns our personal histories and encompasses both episodic and personal semantic memory 4 , 5.
This type of personal semantics, operationalized as autobiographical knowledge or information extracted from repeated autobiographical events, has recently garnered a lot of attention and is thought to be an intermediate entity between semantic and episodic memory 6. The recollective experience results only when details of a specific event are reinstated 5 , 6. Therefore, everyday acts of memory involve guidance by retrieval of personal semantic knowledge culminating in the retrieval of a specific episode 7 , 8 , 9.
Additionally, vivid reminiscence is a hallmark of episodic recollection 10 , 11 and therefore, in this study, we investigate the brain networks that subserve personal semantics and identify the specific parts of these networks that support the phenomenological experience of vivid autobiographical memory. Given the special status of the self in autobiographical memory, it is likely to engage brain networks that have previously been found to be involved in processing information in relation to the self Specifically, the default mode network DMN 13 , 14 has been associated with internally oriented processing across domains like memory 15 , 16 , 17 , 18 , 19 , prospection 20 , 21 , 22 , mental imagery 15 , and mind-wandering Consistent with this general conception of the DMN, an emerging body of neuroimaging work suggests that the human posteromedial cortex, which includes core regions of the DMN such as the retrosplenial cortex, posterior cingulate cortex PCC and the precuneus, is involved in episodic memory 24 , 25 , 26 , 27 , 28 , 29 , Recently, attempts have been made to characterize the various subsystems of the DMN.
Some prior work has investigated levels of activity recruited by personalized image cues versus generic cues 33 but their results do not speak to multivariate representations of content.
It is also not known if retrieving memories of real-world experiences spanning several weeks using highly personalised visual memory cues utilizes the same networks previously identified using generic memory cues in order to represent the content of retrieved memories e. Whereas previous studies compared retrieval of controlled autobiographical memories of pictures taken on campus with retrieval of laboratory events 35 , the current study focuses on naturally occurring autobiographical events extending over much longer spatiotemporal scales with richer personally-relevant attributes.
Recent studies have employed wearable cameras to investigate distributed brain activity patterns during memory retrieval 36 , 37 but they focused on classifying mnemonic output e. Therefore, critical questions remain about the specific functional roles and information content of the various regions of the recollection network 38 , particularly in a relatively more ecologically valid autobiographical reminiscence task. In a previous study focused on the MTL, we found that the anterior hippocampus represents objective space and time content, i.
In the current paper, we perform multivariate pattern analysis on activity across the whole brain to investigate the brain networks that subserve subjective contents i. Participants were recruited using advertisements placed on notice boards in multiple buildings on the main campus of The Ohio State University. To join the study, potential participants had to be willing to participate in the lifelogging data collection and to be willing and able to undergo an MRI scan.
The tenth participant wore the smartphone for 2 weeks. One participant male did not complete the fMRI session due to discomfort in the scanner; therefore, we did not include the data for that participant in any of our analyses. These data were initially collected and analyzed for a previous publication focused on the representation of objective space and time in the MTL 3. Therefore, the task is episodic in nature in the current study as well but the whole-brain multivariate analysis here probes the representation of personal semantic labels of experienced real-world events.