Distinct neural correlates of associative working memory and long-term memory encoding in the medial temporal lobe

Increasing evidence suggests a role for the hippocampus not only in long-term memory (LTM) but also in relational working memory (WM) processes, challenging the view of the hippocampus as being solely involved in episodic LTM. However, hippocampal involvement reported in some neuroimaging studies using "classical" WM tasks may at least partly reflect incidental LTM encoding. To disentangle WM processing and LTM formation we administered a delayed-match-to-sample associative WM task in an event-related fMRI study design. Each trial of the WM task consisted of four pairs of faces and houses, which had to be maintained during a delay of 10s. This was followed by a probe phase consisting of three consecutively presented pairs; for each pair participants were to indicate whether it matched one of the pairs of the encoding phase. After scanning, an unexpected recognition-memory (LTM) task was administered. Brain activity during encoding was analyzed based on WM and LTM performance. Hence, encoding-related activity predicting WM success in the absence of successful LTM formation could be isolated. Furthermore, regions critical for successful LTM formation for pairs previously correctly processed in WM were analyzed. Results showed that the left parahippocampal gyrus including the fusiform gyrus predicted subsequent accuracy on WM decisions. The right anterior hippocampus and left inferior frontal gyrus, in contrast, predicted successful LTM for pairs that were previously correctly classified in the WM task. Our results suggest that brain regions associated with higher-level visuo-perceptual processing are involved in successful associative WM encoding, whereas the anterior hippocampus and left inferior frontal gyrus are involved in successful LTM formation during incidental encoding.     

Encoding activity in anterior medial temporal lobe supports subsequent associative recognition

The ability to bind information together, such as linking a name with a face or a car with a parking space, is a vital process in human episodic memory. To identify the neural bases for this binding process, we measured brain activity during a verbal associative encoding task using event-related functional MRI (fMRI), followed by an associative recognition test for the studied word pairs. Analysis of the encoding data sorted by the associative recognition accuracy allowed us to isolate regions involved in successfully creating associations. We found that encoding activity in bilateral anterior medial temporal lobe (MTL) regions was greater for successfully bound pairs, that is, those later recognized as intact, than for all other pairs. These findings provide evidence that the anterior medial temporal lobes support the successful binding of information in memory.     

Prefrontal and hippocampal contributions to the generation and binding of semantic associations during successful encoding

The ability to form and bind associations between items is an important aspect of successful memory formation. We hypothesize that, during encoding, the left inferior frontal gyrus (IFG) supports generation of associations between items and the hippocampus then binds these associations. This study examined the parametric responses of these regions to varying amounts of generative and relational processing during successful encoding (i.e., for subsequently recognized items). Encoding involved presentation of word triads varying in the number of semantic associations among them (none, one or all); participants judged how many associations were present in each triad. Thus, triads with fewer associations had higher generative load while triads with more associations had higher relational load. Participants later completed a forced-choice recognition test for encoding triads. Successful encoding relative to a control task resulted in activation of bilateral IFG and left hippocampus, and the hippocampus also exhibited a significant subsequent memory effect (hits>misses). Linear parametric analyses revealed that generative load modulated activity in bilateral IFG while relational load correlated with activity in left hippocampus. Although univariate analyses distinguished IFG and hippocampal contributions to the generative and relational stages of encoding, respectively, effective connectivity between these regions did not differ according to condition. Furthermore, this analysis revealed that the left IFG played a pivotal role in coordinating associative encoding processes. Our findings illustrate that modulation of components in a memory network can be independent of patterns of mutual connectivity among those components in mediating successful encoding.     

Dissociable roles of the bilateral anterior temporal lobe in face-name associations: an event-related fMRI study

Previous studies have suggested the importance of bilateral anterior temporal regions in face-name associations, but there is little evidence concerning their precise role. In this fMRI study, we investigated the effects of person-related semantics (PS) and repeated learning (R) on activations in these regions during the retrieval of face-name associations. For encoding stimuli, we prepared four lists of faces. To control the factor of PS, people's and occupation names were attached to the faces in lists A and B, whereas only people's names were attached to the faces in lists C and D. To control the factor of R, the stimuli in lists A and C were learned twice, whereas the stimuli in lists B and D were learned seven times during encoding before fMRI. During fMRI after the encoding, subjects participated in the retrieval task of people's names from faces or in the retrieval task of faces from people's names. The left anterior temporal lobe was significantly activated during the retrieval of people's names from faces encoded with, compared to without, PS; whereas the right anterior temporal lobe was activated during the retrieval of people's faces from names, compared to without, this encoding. Also, activation of the left (but not the right) anterior temporal lobe was significantly reduced after R of face-name associations. These findings suggest that the three components of faces, names, and PS may be mutually mediated by the bilateral anterior temporal lobe, whose activity may be dynamically changed by the level of consolidation of face-name associations.     

Hippocampal functional magnetic resonance imaging during a face-name learning task in adolescents with antecedents of prematurity

Functional magnetic resonance imaging (fMRI) was used to map hippocampal activation during a declarative memory task in a sample of 14 adolescents with antecedents of prematurity (AP). The sample with AP was matched by age, sex and handedness with 14 full-term controls with no history of neurological or psychiatric illness. The target task consisted in learning 16 novel face-name pairs, and the control task involved the examination of two repeated face-name pairs. Stereological methods were also used to quantify hippocampal volumes. In both groups, we observed increased activation in the learning condition compared to the control task in the right fusiform gyrus and the left inferior occipital gyrus, but only premature subjects activated the hippocampus. Group comparison of the activation versus control conditions showed that prematures had greater activity in the right hippocampus than controls during the encoding of the word-face association. Volumetric analyses showed a significant left hippocampal volume loss in adolescents with AP. In addition, we found a significant positive correlation in the premature group between right hippocampal activation and face-name recognition. Functional MRI data also correlated with structural MRI data: right hippocampal activation correlated positively with right hippocampal volume. Our findings are consistent with previous studies of brain plasticity after focal lesions. Left hippocampal tissue loss may be related to an increase in contralateral brain activity, probably reflecting a compensatory mechanism. Our data also suggest that this plasticity is not enough to achieve normal performance.     

Acute effects of alcohol on neural correlates of episodic memory encoding

Although it is well established that alcohol impairs episodic memory encoding, it is unknown how this occurs on a cerebral level. We scanned intoxicated and sober individuals with functional magnetic resonance imaging (fMRI) while they encoded various materials that were recalled the following day. Alcohol impaired memory for object pairs and face-name pairs, but not for words and phrase-word pairs. Impaired performance was associated with reduced bilateral prefrontal activation and non-specific activation of the parahippocampal gyrus. These results suggest that alcohol impairs episodic memory by interfering with activity of regions involved in encoding, and further indicate which regions are critical for human memory.     

Dissociating intentional learning from relative novelty responses in the medial temporal lobe

The establishment of a role for medial temporal lobe (MTL) structures in episodic memory has led to an investigative focus on the specific contributions and interactions between constituent MTL regions, including the hippocampus and surrounding medial temporal cortices. By dissociating an intentional stimulus-category learning condition from a passive viewing condition, we demonstrate, using fMRI, that novelty- and familiarity-driven responses in human anterior and posterior hippocampus, respectively, only occur during intentional learning. With increasing familiarity of stimulus-category associations, there is a shift in neuronal responses from anterior to posterior hippocampal regions. This anterior/posterior response gradient may reflect a weighting of functional hippocampal architecture related to encoding of novel and retrieval of familiar information. By contrast, perirhinal cortex is engaged by novel stimuli irrespective of task, highlighting this region as a component of a generic familiarity discrimination system. By introducing distinct stimulus types, we further demonstrate that these MTL responses are independent of stimulus complexity. Different patterns of activity for intentional learning vs. passive viewing indicate that intentional encoding/retrieval of stimulus-category associations and automatic novelty/familiarity assessment of stimuli are processed in anatomically dissociable neuronal ensembles within the MTL memory system.     

Emotion suppression reduces hippocampal activity during successful memory encoding

People suppressing their emotions while facing an emotional event typically remember it less well. However, the neural mechanisms underlying the impairing effect of emotion suppression on successful memory encoding are not well understood. Because successful memory encoding relies on the hippocampus and the amygdala, we hypothesized that memory impairments due to emotion suppression are associated with down-regulated activity in these brain areas. 59 healthy females were instructed either to simply watch the pictures or to down-regulate their emotions by using a response-focused emotion suppression strategy. Brain activity was recorded using functional magnetic resonance imaging (fMRI), and free recall of pictures was tested afterwards. As expected, suppressing one's emotions resulted in impaired recall of the pictures. On the neural level, the memory impairments were associated with reduced activity in the right hippocampus during successful encoding. No significant effects were observed in the amygdala. In addition, functional connectivity between the hippocampus and the right dorsolateral prefrontal cortex was strongly reduced during emotion suppression, and these reductions predicted free-recall performance. Our results indicate that emotion suppression interferes with memory encoding on the hippocampal level, possibly by decoupling hippocampal and prefrontal encoding processes, suggesting that response-focused emotion suppression might be an adaptive strategy for impairing hippocampal memory formation in highly arousing situations.     

Sleep spindle-related reactivation of category-specific cortical regions after learning face-scene associations

Newly acquired declarative memory traces are believed to be reactivated during NonREM sleep to promote their hippocampo-neocortical transfer for long-term storage. Yet it remains a major challenge to unravel the underlying neuronal mechanisms. Using simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) recordings in humans, we show that sleep spindles play a key role in the reactivation of memory-related neocortical representations. On separate days, participants either learned face-scene associations or performed a visuomotor control task. Spindle-coupled reactivation of brain regions representing the specific task stimuli was traced during subsequent NonREM sleep with EEG-informed fMRI. Relative to the control task, learning face-scene associations triggered a stronger combined activation of neocortical and hippocampal regions during subsequent sleep. Notably, reactivation did not only occur in temporal synchrony with spindle events but was tuned by ongoing variations in spindle amplitude. These learning-related increases in spindle-coupled neocortical activity were topographically specific because reactivation was restricted to the face- and scene-selective visual cortical areas previously activated during pre-sleep learning. Spindle-coupled hippocampal activation was stronger the better the participant had performed at prior learning. These results are in agreement with the notion that sleep spindles orchestrate the reactivation of new hippocampal-neocortical memories during sleep.     

Slow EEG rhythms and inter-hemispheric synchronization across sleep and wakefulness in the human hippocampus

Converging data that attribute a central role to sleep in memory consolidation have increased the interest to understand the characteristics of the hippocampal sleep and their relations with the processing of new information. Neural synchronization between different brain regions is thought to be implicated in long-term memory consolidation by facilitating neural communication and by promoting neural plasticity. However, the majority of studies have focused their interest on intra-hippocampal, rhinal-hippocampal or cortico-hippocampal synchronization, while inter-hemispheric synchronization has been so far neglected. To clarify the features of spontaneous human hippocampal activity and to investigate inter-hemispheric hippocampal synchronization across vigilance states, pre-sleep wakefulness and nighttime sleep were recorded from right and left homologous hippocampal loci using stereo-EEG techniques. Hence, quantitative and inter-hemispheric coherence analyses of hippocampal activity across sleep and waking states were carried out. The results showed the presence of delta activity in human hippocampal spontaneous EEG also during wakefulness. The activity in the delta range exhibited a peculiar bimodal distribution, namely a low frequency non-oscillatory activity (up to 2 Hz) synchronized between hemispheres mainly during wake and REM sleep, and a faster oscillatory rhythm (2-4 Hz). The latter was less synchronized between the hippocampi and seemed reminiscent of animal RSA (rhythmic slow activity). Notably, the low-delta activity showed high inter-hemispheric hippocampal coherence during REM sleep and, to a lesser extent, during wakefulness, paralleled by a (unexpected) decrease of coherence during NREM sleep. Therefore, low-delta hippocampal state-dependent synchronization starkly contrasts with neocortical behavior in the same frequency range. Further studies might shed light on the role of these low frequency rhythms in the encoding processes during wakefulness and in the consolidation processes during subsequent sleep.     

How mood challenges emotional memory formation: an fMRI investigation

Experimental mood manipulations and functional magnetic resonance imaging (fMRI) provide a unique opportunity for examining the neural correlates of mood-congruent memory formation. While prior studies in mood-disorder patients point to the medial temporal lobe in the genesis of mood-congruent memory (MCM) bias, the interaction between mood and emotional memory formation has not been investigated in healthy participants. In particular it remains unclear how regulatory structures in the pre-frontal cortex may be involved in mediating this phenomenon. In this study, event-related fMRI was performed on 20 healthy participants using a full-factorial, within-subjects repeated-measures design to examine how happy and sad moods impact memory for valenced stimuli (positive, negative and neutral words). Main effects of mood, stimulus valence and memory were examined as was activity related to successful memory formation during congruent and in-congruent moods. Behavioral results confirm an MCM bias while imaging results show amygdala and hippocampal engagement in a global mood and successful recall, respectively. MCM formation was characterized by increased activity during mood-congruent encoding of negative words in the orbito-frontal cortex (OFC) and for mood-incongruent processing of negative words in medial- and inferior-frontal gyri (MFG/IFG). These findings indicate that different pre-frontal regions facilitate mood-congruent and incongruent encoding of successfully recalled negative words at the time of learning, with OFC enhancing congruency and the left IFG and MFG helping overcome semantic incongruities between mood and stimulus valence.     

Deficits in learning and memory: parahippocampal hyperactivity and frontocortical hypoactivity in cannabis users

The consumption of cannabis has been linked to impairments in human learning and memory, as well as aspects of executive functioning. Cannabis-related impairments in learning and memory in chronic cannabis users, it has been argued, are caused by the effects of cannabis on hippocampal functioning. The current study involved two experiments. Experiment 1 compared 35 current users of cannabis and 38 well-matched controls on a face-name task, previously shown to activate the hippocampal region. Based on the results of experiment 1, experiment 2 used fMRI and a modified version of the face-name task, to examine cortical and (para)hippocampal activity during learning and recall in 14 current users of cannabis and 14 controls. Results of experiment 1 showed that cannabis users were significantly worse with respect to learning, short and long-term memory performance. Experiment 2 showed that despite non-significant differences in learning and memory performance, cannabis users had significantly lower levels of BOLD activity in the right superior temporal gyrus, right superior frontal gyrus, right middle frontal gyrus and left superior frontal gyrus compared to controls during learning. Results also showed that cannabis users had significantly higher BOLD activity in the right parahippocampal gyrus during learning. Hypoactivity in frontal and temporal cortices, and relative hyperactivity in the parahippocampus identify functional deficits and compensatory processes in cannabis users.     

An examination of the effects of stimulus type, encoding task, and functional connectivity on the role of right prefrontal cortex in recognition memory

Right anterior prefrontal cortex and other brain areas are active during memory retrieval but the role of prefrontal cortex and how it interacts with these other regions to mediate memory function remain unclear. To explore these issues we used positron emission tomography to examine the effects of stimulus material and encoding task on brain activity during visual recognition, assessing both task-related changes and functional connectivity. Words and pictures of objects were encoded using perceptual and semantic strategies, resulting in better memory for semantically encoded items. There was no significant effect of prior encoding strategy on brain activity during recognition. Right anterior prefrontal cortex was equally active during recognition of both types of stimuli irrespective of initial encoding strategy. Regions whose activity was positively correlated with activity in right anterior prefrontal cortex included widespread areas of prefrontal and inferior temporal cortices bilaterally. Activity in this entire network of regions was negatively correlated with recognition accuracy of semantically encoded items. These results suggest that initial encoding task has little impact on the set of brain regions that is active during subsequent recognition. Right anterior prefrontal cortex appears to be involved in retrieval mode, reflected in its equivalent activity across conditions differing in both stimulus type and encoding task, and also in retrieval effort, shown by the negative correlation between its functional connectivity and individual differences in recognition accuracy.     

Neural correlates of successful encoding of semantically and phonologically mediated inter-item associations

This experiment investigated whether the neural correlates of inter-item associative encoding vary according to study task. At study, pairs of unrelated words were subjected to either semantic or phonological relational judgments. Test items comprised studied word pairs (intact), pairs comprised of words belonging to different study pairs (rearranged), and novel pairs. The test requirement was to discriminate between these different classes of test item. fMRI was employed to contrast the neural activity elicited by studied pairs that were correctly endorsed as intact on the later associative recognition test, as opposed to pairs for which associative information was unavailable. In contrast to prior findings for the encoding of single items, there was no evidence that the loci of subsequent associative memory effects varied according to study task. Instead, in both tasks, pairs that were later correctly endorsed as intact elicited enhanced activity in mid- and ventral regions of the left ventrolateral prefrontal cortex (VLPFC). These findings were accompanied by extensive task-invariant reversed subsequent memory effects in medial and lateral parietal and frontal cortices. The findings suggest that the left VLPFC may play a domain-general role in the encoding of item-item associations, and in addition highlight the importance of elucidating the functional significance of reversed subsequent memory effects.     

The hippocampal region is involved in successful recognition of both remote and recent famous faces

There is currently a debate regarding the precise role of medial temporal regions in memory, in particular regarding the time scale of their involvement in conscious recollection of information stored in long-term memory. Using event-related fMRI, we have attempted to contribute to this debate by identifying brain regions associated with the successful recognition of famous faces from two different periods: "Old" faces of people who became famous in the 1960s-1970s and "Recent" faces of people who became famous in the 1990s. We demonstrate that the hippocampus is involved in the successful recognition of famous faces from both periods and does not appear to distinguish between these two periods. We also highlight a network of brain regions, including the left prefrontal cortex, the retrosplenial cortex, the temporo-parietal junction, the caudate and the right cerebellum, which is activated in association with successful recognition of famous faces. Finally, an analysis of the results obtained during a post hoc episodic recognition task shows the specific involvement of anterior hippocampus in the successful encoding of the unfamiliar faces, which were presented during the fame decision task, suggesting a functional distinction between anterior and posterior parts of the hippocampus, the former being specifically involved in successful episodic encoding and the latter being associated with successful retrieval of semantic information.     

Cognitive training-related changes in hippocampal activity associated with recollection in older adults

Impairments in the ability to recollect specific details of personally experienced events are one of the main cognitive changes associated with aging. Cognitive training can improve older adults' recollection. However, little is currently known regarding the neural correlates of these training-related changes in recollection. Prior research suggests that the hippocampus plays a central role in supporting recollection in young and older adults, and that age-related changes in hippocampal function may lead to age-related changes in recollection. The present study investigated whether cognitive training-related increases in older adults' recollection are associated with changes in their hippocampal activity during memory retrieval. Older adults' hippocampal activity during retrieval was examined before and after they were trained to use semantic encoding strategies to intentionally encode words. Training-related changes in recollection were positively correlated with training-related changes in activity for old words in the hippocampus bilaterally. Positive correlations were also found between training-related changes in activity in prefrontal and left lateral temporal regions associated with self-initiated semantic strategy use during encoding and training-related changes in right hippocampal activity associated with recollection during retrieval. These results suggest that cognitive training-related improvements in older adults' recollection can be supported by changes in their hippocampal activity during retrieval. They also suggest that age differences in cognitive processes engaged during encoding are a significant contributor to age differences in recollection during retrieval.     

Phase-locking within human mediotemporal lobe predicts memory formation

Lesion and imaging studies have demonstrated that encoding of declarative memories, i.e. consciously accessible events and facts, is supported by processes within the rhinal cortex and the hippocampus, two substructures of the mediotemporal lobe (MTL). Successful memory formation has, for instance, been shown to be accompanied by the rhinal N400 component, followed by a hippocampal positivity, as well as by transient rhinal-hippocampal phase synchronization. However, it has been an open question, which mediotemporal electroencephalogram (EEG) measures predict memory formation most accurately. Therefore, we analyzed and compared the association of different mediotemporal EEG measures with successful memory formation. EEG characteristics were extracted from intracranial rhinal and hippocampal depth recordings in 31 epilepsy patients performing a continuous word recognition paradigm. Classical event-related potential measures, rhinal-hippocampal synchronization, as well as inter-trial phase-locking and power changes within rhinal cortex and hippocampus were evaluated. We found that inter-trial phase-locking is superior to other EEG measures in predicting subsequent memory. This means that memory formation is related to the precise timing of EEG phases within the MTL with respect to stimulus onset. In particular, early rhinal and hippocampal phase-locking in the alpha/beta range reaching its maximum already between 100 and 300 ms after stimulus onset appears to be a precursor of successful memory formation. Our data suggest that early mediotemporal phase adjustments constitute a relevant mechanism underlying declarative memory encoding.     

When less means more: deactivations during encoding that predict subsequent memory

In event-related functional MRI (fMRI) studies, greater activity for items that are subsequently remembered (R-items) than for items that are subsequently forgotten (F-items), or Dm effect (Difference in memory), has been attributed to successful encoding operations. In contrast, regions showing a reverse DM effect (revDM = F-items > R-items) have been linked to detrimental processes leading to forgetting. Yet, revDMs may reflect not only activations for F-items (aFs) but also deactivations for R-items (dRs), and the latter alternative is more likely to reflect beneficial rather than detrimental encoding processes. To investigate this issue, we used a paradigm that included a fixation baseline and could distinguish between the two types of revDMs (aF vs. dR). Participants were scanned while encoding semantic associations between words or perceptual associations between words and fonts, and their memory was measured with associative recognition tests. For both semantic and perceptual encoding, dR effects were found in dorsolateral prefrontal, temporoparietal, and posterior midline regions. In contrast with a prior study that attributed revDMs in these regions to detrimental processes, the present results suggest that these effects reflect beneficial processes, that is, the efficient reallocation of neurocognitive resources. At the same time, aF effects were found in other regions, such as the insula, and these are more consistent with an interpretation in terms of detrimental processes. Whereas most fMRI studies of encoding have focused on activation increases, the present study indicates that activation decreases are also critical for successful learning of new information.     

Evidence for functional specialization of hippocampal subfields detected by MR subfield volumetry on high resolution images at 4 T

Animal studies suggest an involvement of CA3 and dentate gyrus (CA3&DG) in memory encoding and early retrieval and an involvement of CA1 in late retrieval, consolidation and recognition. The aim of this study was to test if similar associations could be found between hippocampal subfield volumes measured in vivo using a manual parcellation scheme and selected scores of the California Verbal Learning Test II (CVLTII): total immediate free recall discriminability (IFRD), short free recall discriminability (SFRD), and delayed recall discriminability (DRD). 50 elderly subjects (25 controls and 25 cognitively impaired subjects) had CVLTII and high resolution hippocampal MRI at 4T. Entorhinal cortex, subiculum, CA1, CA1-CA2 transition zone, and CA3&DG were manually marked on five slices in the anterior hippocampal body on the MRI. Pearson correlations followed by stepwise regression analysis were used to test for associations between subfield volumes and CVLTII. IFRD and SFRD, which are measures of encoding/early retrieval, were associated with CA3&DG, and DRD, which measures consolidation/late retrieval, with CA1. These preliminary findings demonstrate that subfield volumetry has the potential to study non invasively subfield specific memory functions.     

fMRI BOLD response to increasing task difficulty during successful paired associates learning

We used functional magnetic resonance imaging (fMRI) to assess cortical activations associated with increasing task difficulty (TD) in a visuospatial paired associates learning task. Encoding and retrieval were examined when 100% successful retrieval of three, four, or six object-location pairs had been attained (thus ensuring that performance was matched across subjects). As memory load increased, in general, the number of attempts taken to achieve 100% successful retrieval increased, while the number of trials correctly completed on the first attempt decreased. By modelling parametric variations in working memory load with BOLD signal changes we were able to identify brain regions displaying linear and nonlinear responses to increasing load. During encoding, load-independent activations were found in occipitoparietal cortices (excluding the precuneus for which linear load dependency was demonstrated), anterior cingulate, and cerebellum, while linear load-dependent activations in these same regions were found during retrieval. Nonlinear load-dependent responses, as identified by categorical contrasts between levels of load, were found in the right DLPFC and left inferior frontal gyrus. The cortical response to increasing cognitive demands or TD appears to involve the same, rather than an additional, network of brain regions "working harder."