Relational learning and transitive expression in aging and amnesia.

Aging has been associated with a decline in relational memory, which is critically supported by the hippocampus. By adapting the transitivity paradigm (Bunsey and Eichenbaum (1996) Nature 379:255-257), which traditionally has been used in nonhuman animal research, this work examined the extent to which aging is accompanied by deficits in relational learning and flexible expression of relational information. Older adults' performance was additionally contrasted with that of amnesic case DA to understand the critical contributions of the medial temporal lobe, and specifically, the hippocampus, which endures structural and functional changes in healthy aging. Participants were required to select the correct choice item (B versus Y) based on the presented sample item (e.g., A). Pairwise relations must be learned (A->B, B->C, C->D) so that ultimately, the correct relations can be inferred when presented with a novel probe item (A->C?Z?). Participants completed four conditions of transitivity that varied in terms of the degree to which the stimuli and the relations among them were known pre-experimentally. Younger adults, older adults, and DA performed similarly when the condition employed all pre-experimentally known, semantic, relations. Older adults and DA were less accurate than younger adults when all to-be-learned relations were arbitrary. However, accuracy improved for older adults when they could use pre-experimentally known pairwise relations to express understanding of arbitrary relations as indexed through inference judgments. DA could not learn arbitrary relations nor use existing knowledge to support novel inferences. These results suggest that while aging has often been associated with an emerging decline in hippocampal function, prior knowledge can be used to support novel inferences. However, in case DA, significant damage to the hippocampus likely impaired his ability to learn novel relations, while additional damage to ventromedial prefrontal and anterior temporal regions may have resulted in an inability to use prior knowledge to flexibly express indirect relational knowledge.

Intact learning of new relations in amnesia as achieved through unitization.

Hippocampal amnesia is defined by deficits in the binding of relations among items--a deficit captured by the transverse patterning (TP) task. Unitization is a processing mechanism that may allow amnesic patients to compensate for relational memory deficits. Amnesic patient D.A. demonstrated intact TP, and performance was maintained 1 month following training. Successful acquisition of relations occurred only when D.A. fused or integrated objects into a unified representation. D.A. did not acquire relations when he did not generate such integrated scenarios, and acquisition of relations was slowed when integration had to occur for novel stimuli. Amnesic patients K.C. and R.F.R. were tested to provide comparative data; K.C. and R.F.R. did not benefit from unitization, perhaps due to additional cortical damage. We propose that unitization requires visual imagery of multiple items that are fused/integrated; through the benefit of extended on-line maintenance, this fused representation is anchored to existing representations in semantic memory.

Hippocampal lateralization and memory in children and adults.

The neural organization of cognitive processes, particularly hemispheric lateralization, changes throughout childhood and adolescence. Differences in the neural basis of relational memory between children and adults are not well characterized. In this study we used magnetoencephalography to observe the lateralization differences of hippocampal activation in children and adults during performance of a relational memory task, transverse patterning (TP). The TP task was paired with an elemental control task, which does not depend upon the hippocampus. We contrasted two hypotheses; the compensation hypothesis would suggest that more bilateral activation in children would lead to better TP performance, whereas the maturation hypothesis would predict that a more adult-like right-lateralized pattern of hippocampal activation would lead to better performance. Mean-centered partial least squares analysis was used to determine unique patterns of brain activation specific to each task per group, while diminishing activation that is consistent across tasks. Our findings support the maturation hypothesis that a more adult-like pattern of increased right hippocampal lateralization in children leads to superior performance on the TP task. We also found dynamic changes of lateralization throughout the time course for all three groups, suggesting that caution is needed when interpreting conclusions about brain lateralization.

Techniques for detection and localization of weak hippocampal and medial frontal sources using beamformers in MEG.

Magnetoencephalography provides precise information about the temporal dynamics of brain activation and is an ideal tool for investigating rapid cognitive processing. However, in many cognitive paradigms visual stimuli are used, which evoke strong brain responses (typically 40-100 nAm in V1) that may impede the detection of weaker activations of interest. This is particularly a concern when beamformer algorithms are used for source analysis, due to artefacts such as "leakage" of activation from the primary visual sources into other regions. We have previously shown (Quraan et al. 2011) that we can effectively reduce leakage patterns and detect weak hippocampal sources by subtracting the functional images derived from the experimental task and a control task with similar stimulus parameters. In this study we assess the performance of three different subtraction techniques. In the first technique we follow the same post-localization subtraction procedures as in our previous work. In the second and third techniques, we subtract the sensor data obtained from the experimental and control paradigms prior to source localization. Using simulated signals embedded in real data, we show that when beamformers are used, subtraction prior to source localization allows for the detection of weaker sources and higher localization accuracy. The improvement in localization accuracy exceeded 10 mm at low signal-to-noise ratios, and sources down to below 5 nAm were detected. We applied our techniques to empirical data acquired with two different paradigms designed to evoke hippocampal and frontal activations, and demonstrated our ability to detect robust activations in both regions with substantial improvements over image subtraction. We conclude that removal of the common-mode dominant sources through data subtraction prior to localization further improves the beamformer's ability to project the n-channel sensor-space data to reveal weak sources of interest and allows more accurate localization.

Detection and localization of hippocampal activity using beamformers with MEG: a detailed investigation using simulations and empirical data.

The ability to detect neuronal activity emanating from deep brain structures such as the hippocampus using magnetoencephalography has been debated in the literature. While a significant number of recent publications reported activations from deep brain structures, others reported their inability to detect such activity even when other detection modalities confirmed its presence. In this article, we relied on realistic simulations to show that both sides of this debate are correct and that these findings are reconcilable. We show that the ability to detect such activations in evoked responses depends on the signal strength, the amount of brain noise background, the experimental design parameters, and the methodology used to detect them. Furthermore, we show that small signal strengths require contrasts with control conditions to be detected, particularly in the presence of strong brain noise backgrounds. We focus on one localization technique, the adaptive spatial filter (beamformer), and examine its strengths and weaknesses in reconstructing hippocampal activations, in the presence of other strong brain sources such as visual activations, and compare the performance of the vector and scalar beamformers under such conditions. We show that although a weight-normalized beamformer combined with a multisphere head model is not biased in the presence of uncorrelated random noise, it can be significantly biased in the presence of correlated brain noise. Furthermore, we show that the vector beamformer performs significantly better than the scalar under such conditions. We corroborate our findings empirically using real data and demonstrate our ability to detect and localize such sources.

Learning related activation of somatosensory cortex by an auditory stimulus recorded with magnetoencephalography.

Advances in non-invasive neuroimaging technology now provide a means of directly observing learning within the brain. Classical conditioning serves as an ideal starting point for examining the dynamic expression of learning within the human brain, since this paradigm is well characterized using multiple levels of analysis in a broad range of species. We used MEG to expand the characterization of conditioned responses (CR) recorded from the human brain with a simultaneous examination of their spatial, temporal and spectral properties. We paired an auditory conditioned stimulus (CS+) with a somatosensory unconditioned stimulus (US). We found that when the US was randomly omitted, presentations of CS+ alone, elicited greater desynchronization of beta-band activity in contralateral somatosensory cortex compared to presentations of an auditory stimulus that was never paired with the US (CS-), and compared the CS+ following a non-reinforced extinction session. This differentiation was largest between 150 and 350ms following US omission. We show that cross-modal CRs in the primary sensorimotor system are predominantly characterized by modulation of ongoing cortical oscillations.

Neural system interactions underlying human transitive inference.

Human problem solving relies on multiple strategies supported by dynamic neural network interactions. The transitive inference (TI) problem solving task can be accomplished by the extraction of relations among stimuli or by responding to reinforcement histories of items using associative learning. Relational and associative strategies are assumed to rely on the hippocampus and caudate nucleus, respectively; which compete to control behavior. However, we found that increased recruitment of both systems in TI is correlated with greater accuracy and awareness, and reduced associative responding to single items. Contrary to prior assumptions, the hippocampus and caudate interact cooperatively to facilitate successful TI. We suggest that the dynamics of the relationship between the hippocampus and caudate depends critically upon task demands.

Relational framework improves transitive inference across age groups.

Transitive inference is a complex task, conducive to the use of multiple strategies. We investigated whether transitive inference accuracy can be improved by biasing strategy choice towards a proposition-based approach that relies on the extraction of relations among stimuli. We biased strategy choice by using familiar stimuli with known relations that tap prior knowledge. Semantic information led to increased accuracy for younger and older adults, and increased awareness of stimulus relations. Increased age was associated with reduced awareness. Awareness accounted for the variability in performance accuracy to a greater extent than age, as aware older and younger adults showed similar accuracies on all conditions. The current work indicates that age differences in performance can be minimized by providing semantically meaningful stimuli that bias participants to use a relational proposition-based approach.

A complementary analytic approach to examining medial temporal lobe sources using magnetoencephalography.

Neuropsychological and neuroimaging findings reveal that the hippocampus is important for recognition memory. However, it is unclear when and whether the hippocampus contributes differentially to recognition of previously studied items (old) versus novel items (new), or contributes to a general processing requirement that is necessary for recognition of both types of information. To address this issue, we examined the temporal dynamics and spectral frequency underlying hippocampal activity during recognition of old/new complex scenes using magnetoencephalography (MEG). In order to provide converging evidence to existing literature in support of the potential of MEG to localize the hippocampus, we reconstructed brain source activity using the beamformer method and analyzed three types of processing-related signal changes by applying three different analysis methods: (1) Synthetic aperture magnetometry (SAM) revealed event related and non-event-related spectral power changes; (2) Inter-trial coherence (ITC) revealed time-locked changes in neural synchrony; and (3) Event-related SAM (ER-SAM) revealed averaged event-related responses over time. Hippocampal activity was evident for both old and new information within the theta frequency band and during the first 250 ms following stimulus onset. The early onset of hippocampal responses suggests that general comparison processes related to recognition of new/old information may occur obligatorily.

Semantic information alters neural activation during transverse patterning performance.

Memory tasks can be performed using multiple cognitive strategies, which are mediated by different brain systems. The transverse patterning (TP) task is dependent upon the integrity of the hippocampal system, however, we previously demonstrated successful TP following hippocampal damage using meaningful stimuli and relations (Moses, S.N., Ostreicher, M.L., Rosenbaum, R.S., Ryan, J.D., 2008. Successful transverse patterning in amnesia using semantic knowledge. Hippocampus 18, 121-124). Here, we used magnetoencephalgraphy (MEG) to directly observe the neural underpinnings of TP, and the changes that occur as stimuli and relations become more meaningful. In order to optimize our ability to detect signal from deep, non-dominant, brain sources we implemented the event-related synthetic aperture magnetometry minimum-variance beamformer algorithm (ER-SAM; Cheyne, D., Bakhtazad, L., Gaetz, W., 2006. Spatiotemporal mapping of cortical activity accompanying voluntary movements using an event-related beamforming approach. Human Brain Mapping 27, 213-229) coupled with the partial least squares (PLS) multivariate statistical approach (McIntosh, A.R., Bookstein, F.L., Haxby, J.V., Grady, C.L., 1996. Spatial pattern analysis of function brain images using partial least squares. NeuroImage 3, 143-157; McIntosh, A.R., Lobaugh, N.J., 2004. Partial least squares analysis of neuroimaging data: Applications and advances. NeuroImage 23, S250-S263). We found that increased meaningfulness elicited reduced bilateral hippocampal activation, along with increased activation of left prefrontal and temporal cortical structures, including inferior frontal (IFG), as well as anterior temporal and perirhinal cortices. These activation patterns may represent a shift towards reliance upon existing semantic knowledge. This shift likely permits successful TP performance with meaningful stimuli and relations following hippocampal damage.

Impaired relational organization of propositions, but intact transitive inference, in aging: Implications for understanding underlying neural integrity.

The ability to perform relational proposition-based reasoning was assessed in younger and older adults using the transitive inference task in which subjects learned a series of premise pairs (A>B, B>C, C>D, D>E, E>F) and were asked to make inference judgments (B?D, B?E, C?E). Learning of premise pairs was related to subsequent inference performance and conscious awareness of the stimulus hierarchy (A>B>C>D>E>F). Despite extended training, age-related deficits were observed for response times, accurate learning of the premise pairs, making inference judgments, and articulating the hierarchy. When performance for younger and older adults was examined with respect to whether they were subsequently considered aware of the hierarchy, older and younger adults still significantly differed on their accuracy for studied premise pairs, but performance between the age groups was similar for the inference pairs. Successful transitive inference performance is contingent upon the relational organization of propositions within memory and such processes are impaired in aging, potentially leading to disruptions in conscious access to the stimulus hierarchy. Such findings, in concert with previous neuropsychological and neuroimaging studies, implicate an age-related deficit in the functioning of frontal and medial temporal lobe structures, with particular emphasis on the hippocampus.

Hippocampus volume and episodic memory in schizophrenia.

Previous studies of schizophrenia have suggested a linkage between neuropsychological (NP) deficits and hippocampus abnormality. The relationship between hippocampus volume and NP functioning was investigated in 24 patients with chronic schizophrenia and 24 matched healthy controls. Overall intracranial, white and gray matter, and anterior (AH) and posterior (PH) hippocampus volumes were assessed from magnetic resonance images (MRI). NP domains of IQ, attention, and executive function were also evaluated with respect to volumetric measures. It was hypothesized that AH and PH volumes and episodic memory scores would be positively associated in controls and that the schizophrenia group would depart from this normative pattern. NP functioning was impaired overall and AH volume was smaller in the schizophrenia group. In the controls, the hippocampus-memory relationships involved AH and PH, and correlations were significant for verbal memory measures. In the schizophrenia group, positive correlations were constrained to PH. Negative correlations emerged between AH and verbal and visual memory measures. For both groups, cortical volume negatively correlated with age, but a negative correlation between age and hippocampus volume was found only in the schizophrenia group. In this sample of adults with schizophrenia, atypical relationships between regional hippocampus volumes and episodic memory ability were found, as was an atypical negative association between hippocampus volume and age.

Successful transverse patterning in amnesia using semantic knowledge.

We examined whether alternate systems compensate for deficient ones to process relations in amnesia. Transverse patterning (TP), a test of relating items to one another in memory, is reliably impaired in amnesia and has played a central role in testing hippocampal function. We facilitated successful TP in amnesia by providing familiar stimuli with semantically meaningful relationships. Tapping semantic knowledge allowed TP to be solved via extrahippocampal structures. Our work shows that by framing a situation as meaningful we can engage alternate neural systems to compensate for impairments.

Cognitive integrity predicts transitive inference performance bias and success.

Transitive inference has traditionally been regarded as a relational proposition-based reasoning task, however, recent investigations question the validity of this assumption. Although some results support the use of a relational proposition-based approach, other studies find evidence for the use of associative learning. We examined whether participants are biased towards a relational proposition-based approach, or associative learning, depending upon the integrity of other cognitive abilities supported by frontal and medial temporal lobes. We found that transitive inference success and strategy use are related to tasks that require processing of, and memory for, relations. Participants who showed superior memory for relations among items showed superior transitive inference accuracy, and conversely, participants who showed superior memory for single items showed inferior transitive inference accuracy. Participants who demonstrated performance patterns indicative of a proposition-based approach showed more accurate inference performance, and superior memory for, and online access to, relations among items. Participants who demonstrated performance patterns indicative of associative learning showed inferior transitive inference accuracy and superior iconic/pictorial abilities. We speculate that transitive inference performance is mediated by interactions among multiple cognitive systems that support different aspects of processing, from which the degree of contribution varies depending on underlying cognitive/neural integrity.

Dynamic neural activity recorded from human amygdala during fear conditioning using magnetoencephalography.

Magnetoencephalography (MEG) was used to record the dynamics of amygdala neuronal population activity during fear conditioning in human participants. Activation during conditioning training was compared to habituation and extinction sessions. Conditioned stimuli (CS) were visually presented geometric figures, and unconditioned stimuli (US) were aversive white-noise bursts. The CS+ was paired with the US on 50% of presentations and the CS- was never paired. The precise temporal resolution of MEG allowed us to address the issue of whether the amygdala responds to the onset or offset of the CS+, and/or the expectation of the initiation or offset of the an omitted auditory US. Fear conditioning elicited differential amygdala activation for the unpaired CS+ compared to the CS-, extinction and habituation. This was especially robust in the right hemisphere at CS onset. The strongest peaks of amygdala activity occurred at an average of 270 ms in the right and 306 ms in the left hemisphere following unpaired CS+ onset, and following offset at 21 ms in the left and 161 ms in the right (corresponding to an interval of 108 ms and 248 ms after the anticipated onset of the US, respectively). However, the earliest peaks in this epoch preceded US onset in most subjects. Thus, the activity dynamics suggest that the amygdala both differentially responds to stimuli and anticipates the arrival of stimuli based on prior learning of contingencies. The amygdala also shows stimulus omission-related activation that could potentially provide feedback about experienced stimulus contingencies to modify future responding during learning and extinction.

Impaired secondary somatosensory gating in patients with schizophrenia.

A large and growing literature has demonstrated a deficit in auditory gating in patients with schizophrenia. Although that deficit has been interpreted as a general gating problem, no deficit has been shown in other sensory modalities. Recent research in our laboratory has examined sensory gating effects in the somatosensory system showing no difference in gating of the primary somatosensory response between patients with schizophrenia and control subjects. This is consistent with recent structural studies showing no cortical structural abnormality in primary somatosensory area in schizophrenia. However, a significant decrease in cortical thickness and gray matter volume loss in secondary somatosensory cortex has recently been reported, suggesting this as a focus for impaired somatosensory gating. Thus, the current study was designed (1) to replicate previous work showing a lack of schizophrenia deficit in primary somatosensory cortex (SI) gating, and (2) to investigate a possible deficit in secondary somatosensory cortex (SII) gating. In a paired-pulse paradigm, dipolar sources were assessed in SI and SII contralateral to unilateral median nerve stimulation. Patients demonstrated no impairment in SI gating, but a robust gating deficit in SII, supporting the presence of cross modal gating deficits in schizophrenia.

Brain regions and their dynamics in prospective memory retrieval: a MEG study.

We measured brain activity using magnetoencephalography in five participants during ongoing tasks that included prospective memory, retrospective memory, and oddball trials. Sources were identified in the hippocampal formation and posterior parietal and frontal lobes. Posterior parietal cortex activation had an earlier onset in the prospective memory condition than retrospective memory or oddball conditions, a higher level of theta activity in the retrospective condition, and higher levels of upper alpha in the prospective and oddball conditions. Activation of the hippocampal formation had a longer duration in the retrospective memory and prospective memory conditions than the oddball condition, but prominent alpha and theta band activity was present in all three conditions. We interpret the early (87 ms) onset of activity in parietal cortex as evidence for an initial noticing of appropriate conditions for a PM response. Hippocampal activity may reflect a subsequent memory search for the intended action.

An investigation of learning strategy supporting transitive inference performance in humans compared to other species.

Generalizations about neural function are often drawn from non-human animal models to human cognition, however, the assumption of cross-species conservation may sometimes be invalid. Humans may use different strategies mediated by alternative structures, or similar structures may operate differently within the context of the human brain. The transitive inference problem, considered a hallmark of logical reasoning, can be solved by non-human species via associative learning rather than logic. We tested whether humans use similar strategies to other species for transitive inference. Results are crucial for evaluating the validity of widely accepted assumptions of similar neural substrates underlying performance in humans and other animals. Here we show that successful transitive inference in humans is unrelated to use of associative learning strategies and is associated with ability to report the hierarchical relationship among stimuli. Our work stipulates that cross-species generalizations must be interpreted cautiously, since performance on the same task may be mediated by different strategies and/or neural systems.

A specific test of hippocampal deficit in schizophrenia.

Despite numerous studies in which hippocampal abnormalities were found, schizophrenia patients' hippocampal neural activity has not been systematically evaluated on a specific hippocampal-dependent task. The transverse-patterning task (TP) is sensitive to the relational mnemonic capabilities of the hippocampus. Ten schizophrenia patients and 10 controls performed TP and control tasks that are not hippocampal dependent. As predicted, patients displayed a behavioral impairment in TP and not in control tasks. Magnetoencephalography showed controls activating right hippocampus during TP performance. Patients showed more bilateral or left hippocampal activation during TP, and greater left lateralization was associated with better performance on TP. Patients' abnormal hippocampal lateralization may play a role in the hippocampal-dependent behavioral deficit.

M50 sensory gating predicts negative symptoms in schizophrenia.

Impaired auditory sensory gating is considered characteristic of schizophrenia and a marker of the information processing deficit inherent to that disorder. Predominance of negative symptoms also reflects the degree of deficit in schizophrenia and is associated with poorer pre-morbid functioning, lower IQ, and poorer outcomes. However, a consistent relationship between auditory sensory gating and negative symptoms in schizophrenia has yet to be demonstrated. The absence of such a finding is surprising, since both impaired auditory gating and negative symptoms have been linked with impaired fronto-temporal cortical function. The present study measured auditory gating using the P50 event related potential (ERP) in a paired-click paradigm and capitalized on the relative localization advantage of magnetoencephalography (MEG) to assess auditory sensory gating in terms of the event related field (ERF) M50 source dipoles on bilateral superior temporal gyrus (STG). The primary hypothesis was that there would be a positive correlation between lateralized M50 auditory sensory gating measures and negative symptoms in patients with schizophrenia. A standard paired-click paradigm was used during simultaneous EEG and MEG data collection to determine S2/S1 sensory gating ratios in a group of 20 patients for both neuroimaging techniques. Participants were administered the Schedule for the Assessment of Negative Symptoms (SANS), the Positive and Negative Symptom Scale (PANSS), and the Calgary Depression Scale for Schizophrenia. Consistent with previous reports, there was no relationship between ERP P50 sensory gating and negative symptoms. However, right (not left) hemisphere ERF M50 sensory gating ratio was significantly and positively correlated with negative symptoms. This finding is compatible with information processing theories of negative symptoms and with more recent findings of fronto-temporal abnormality in patients with predominantly negative symptoms.