Cortical Network Connectivity Tracks Working Memory Load

Post by Shireen Parimoo

What's the science?

Working memory is the ability to maintain and manipulate information in your mind – like trying to keep directions in mind while navigating. Working memory is critical for a wide range of cognitive functions such as learning, problem-solving, and decision-making, however, it is restricted in its capacity, meaning that only a limited amount of information can be processed in working memory at a given time. Thus, performance suffers when working memory load – or the amount of information to be processed – is higher than the working memory capacity. Interestingly, research shows that the connectivity between different brain regions is consistent across different working memory loads. However, it is possible that conventional analysis techniques do not capture subtle changes in connectivity patterns during working memory tasks. This week in NeuroImage, Eryilmaz, and colleagues used functional magnetic resonance imaging and machine learning techniques to identify load-dependent differences in functional connectivity during a working memory task.

How did they do it?

Participants were 177 healthy adults who completed the Sternberg item recognition task while undergoing fMRI scanning. In the encoding portion of the task, participants were shown a list of 1, 3, 5, or 7 consonants to remember. After a brief delay period, they were sequentially presented with 14 probe letters and had to indicate if they had seen those letters before (targets) or not (foils). Working memory load typically increases with an increasing number of targets (1T-7T) at encoding, resulting in longer response times (RT) at retrieval.

The authors first investigated how the functional connectivity of various brain regions during retrieval changed with increasing working memory load. These brain regions were assigned to one of seven brain networks, such as the dorsal and ventral attention networks (DAN/VAN), default mode network (DMN), and the frontoparietal control network (FPCN). A linear support vector machine (SVM) classifier (machine learning) was used to determine if the different working memory load conditions could be distinguished based on the patterns of functional connectivity within and between different brain networks. The authors further identified the strongest patterns of functional connectivity that could discriminate the lowest (1T) from the highest (7T) working memory load conditions using Neighborhood Component Analysis (a clustering machine learning technique). Finally, they used a leave-one-out cross-validation approach to determine whether connectivity within individual brain networks and across the entire cerebral cortex (global connectivity) predicted behavioral performance.

What did they find?

As expected, participants’ response times increased linearly with increasing working memory load. The load conditions could be reliably decoded from functional connectivity patterns across the brain using the SVM classifier. Classifier accuracy was greater when distinguishing between conditions with larger load differences (e.g., 3T vs. 7T) than between conditions with a smaller difference in working memory load (e.g., 5T vs. 7T). Functional connectivity of brain regions both within and between different brain networks tracked working memory load. For example, within-network connectivity in the VAN and the FPCN most strongly distinguished between the lowest (1T) and highest (7T) load conditions. Similarly, connectivity between regions in the FPCN and the DMN, along with VAN-DMN, DAN-VAN, and VAN-FPCN networks could also be used to distinguish between these load conditions. Finally, differences in connectivity within each individual network was not correlated with behavioral performance. However, a measure of network strength derived from total within-network connectivity as well as from global connectivity during the 1T and 7T conditions successfully predicted the differences in participants’ RT during the task.

What's the impact?

The authors of this study used machine learning techniques to demonstrate how the reconfiguration of connectivity patterns between a distributed network of brain regions varies with working memory load. Moreover, these connectivity differences were also predictive of behavioral performance in healthy adults. This is exciting because this approach can be further extended to improve our understanding of subtle differences in brain network dynamics in neuropsychiatric conditions, such as those characterized by working memory deficits.

Eryilmaz et al. Working memory load-dependent changes in cortical network connectivity estimated by machine learning. NeuroImage (2020). Access the original scientific publication here.

Type I Interferon-Mediated Signaling Produces Nociceptive Pain

Post by Amanda McFarlan 

What's the science?

Viral infections can cause both acute and chronic neuropathic pain. However, the mechanisms by which these neuropathies occur are still poorly understood. Type I interferons (IFNs) are cytokines that are rapidly released by a variety of cells in response to viral infection and therefore, these IFNs may be an ideal candidate for mediating viral-induced neuropathic pain by directly binding nociceptors in the body’s peripheral nervous system. This week in the Journal of Neuroscience, Barragán-Iglesias and colleagues investigated the role of type I IFNs in mediating nociceptive pain responses. 

How did they do it?

The authors used the von Frey filament test (to test mechanical sensitivity) and the Hargreaves test (to test thermal pain sensitivity) to evaluate the nociceptive responses in mice following intraplantar (in the foot) injections of IFN-α, IFN-β or saline. Because they found that IFN administration caused a pain response in mice, the authors used RNAscope in situ hybridization (detects RNA in intact cells) to determine whether IFN receptors could be found in the dorsal root ganglion (a cluster of neurons in the spinal cord that relays sensory information to the brain). To investigate the effect of IFN on downstream signaling in the cell, the authors applied IFN-α and IFN-β to cultured dorsal root ganglion neurons and performed western blots to identify which downstream signaling molecules were present. Since the MNK-eIF4E pathway (involved in nociception) was found to be activated by IFN, the authors performed intraplantar injections of IFN-α and IFN-β in transgenic mice with disrupted MNK-eIF4E pathways to investigate the role of this pathway in IFN-mediated pain responses. Next, the authors used patch-clamp electrophysiology in cultured dorsal root ganglion neurons treated with IFN-α to examine the effects of IFN on neuronal excitability. Finally, the authors investigated the role of endogenous IFN in the pain response. To do this, they simulated a viral infection by injecting poly I:C (an immunostimulant) in transgenic mice with disrupted MNK-eIF4E pathways and wild-type mice and then measured mechanical and thermal pain sensitivity. 

What did they find?

The authors found that administration of IFN-α and IFN-β, but not saline, in mice caused increased hypersensitivity to mechanical stimulation with no change in sensitivity to thermal stimulation. They confirmed that the majority of dorsal root ganglion cells expressed the subunits for IFN receptors, suggesting that IFN-mediated nociceptive responses are communicated to the brain via these receptors. Then, the authors determined that the application of IFN-α and IFN-β quickly activated downstream signaling molecules in the MNK-eIF4E pathway. They found that compared to controls, transgenic mice with disruptions in the MNK-eIF4E pathway had reduced hypersensitivity to mechanical stimulation following IFN administration, suggesting this pathway is critical for IFN-mediated nociceptive pain responses. Next, the authors determined that dorsal root ganglion neurons treated with IFN had increased neuronal excitability compared to controls. Notably, treatment with IFN reduced the latency of action potential initiation, suggesting that IFN-mediated responses in dorsal root ganglion neurons cause rapid neuronal hyperexcitability that coincides with activation of the MNK-eIF4E pathway.

Finally, the authors showed that administration of poly I:C increased mechanical and thermal hypersensitivity in wild-type mice, but not in transgenic mice with disrupted MNK-eIF4E pathways. They found that direct application of poly I:C to dorsal root ganglion neurons did not cause an increase in the signaling molecules in the MNK-eIF4E pathway, which suggests that the in vivo nociceptive response to poly I:C is likely mediated by the endogenous release of IFN that acts on dorsal root ganglion neurons. 

What’s the impact?

This is the first study to provide evidence for a mechanism by which a viral infection, which causes the release of IFN, can rapidly induce nociceptive responses. The authors showed that type 1 IFNs act via the MNK-eIF4E signaling pathway to produce increased mechanical hypersensitivity. Together, these findings highlight the relevance of the IFN to MNK-eIF4E pathway as a potential target for the development of treatments to alleviate viral-induced acute and chronic neuropathic pain. 

Barragán-Iglesias et al. Type I Interferons Act Directly on Nociceptors to Produce Pain Sensitization: Implications for Viral Infection-Induced Pain. The Journal of Neuroscience (2020). Access the original scientific publication here.

Characterizing the Neural Signature of Preferences

Post by Elisa Guma

What's the science?

When we make a decision, typically we identify our options, estimate the value of those options, and compare the values to select the best option. Several neural and computational approaches have been employed to try to understand the valuation process, and the brain networks involved. However, the mechanisms behind decision making remain poorly understood. A few key brain regions have been identified as playing a key role in subjective valuation, also termed ‘brain valuation system’ including the ventromedial prefrontal cortex, ventral striatum, and posterior cingulate cortex, however, several other regions have also been identified as playing a key role. This week in Nature Neuroscience, Lopez-Persem and colleagues use a large dataset of intracranial electrophysiological recordings in humans (being treated with epilepsy) to better identify which brain regions and what type of underlying activity is involved in generating value signals in judgement tasks.

How did they do it?

The authors’ first goal was to identify brain regions in which value signals were detectable during judgement tasks. In other words: to identify the brain valuation system. The authors collected intracranial electroencephalography (iEEG) data from 4,273 intracranial electrodes in the brains of 36 patients being treated for drug-resistant focal epilepsy (across 3 treatment centers). Each participant had between 12-18 electrodes implanted for seizure localization. Participants performed judgments tasks while neural activity was being recorded via electrodes. Some participants performed a short version of the task, and others performed a longer version. The longer version of the task started with a “distracting task” in which participants were asked to estimate the age of faces and paintings, and rate how confident they felt in their guess. In the second phase of the task, participants were asked to rate the likeability of food items as well as faces and paintings, followed by a confidence rating. In a third phase, participants were given a choice of two pictures belonging to the same category (face, food, painting) and asked to choose the one they liked best. In the short version of the task, participants only completed the second and third phases with food item images and were not asked to rate confidence. 

The authors sought to find the time window in which each brain region of interest was most associated with the value signal by investigating the relationship between the subjective value given by participants and parcellated brain activity (77 regions using the Automated Anatomical Labeling atlas). They focused their analyses on high-gamma-band (50-150Hz) activity because it is thought to be a close reflection of local neuron spiking activity. The authors wanted to know if the electrophysiological activity recordings were related to four core properties of subjective valuation. They assessed anticipation with pre-stimulus activity, generality with the inclusion of non-food items, automaticity by including other types of rating (age), and quadratic coding by measuring confidence in ratings.

What did they find?

The authors identified a quadratic, or U-shaped relationship between first-order ratings (age or likeability of an item) and second-order (confidence in those ratings) ratings. This was true for both food and non-food items as well as age and likeability tasks. The authors also observed that the likeability ratings were reliable estimates of subjective value, able to predict choice, reaction time, and confidence. The authors identified 18 regions of interest (among the 77 analyzed) in which a significant subjective value representation was associated. The authors identified several regions belonging to the brain valuation system: the orbitofrontal cortex OFC (comprising the ventromedial prefrontal cortex and lateral orbitofrontal cortex), and parahippocampal complex (PHC) comprising the hippocampus and parahippocampal cortex. They also found significant associations with other regions including the anterior cingulate gyrus, fusiform area, inferior temporal cortex, and inferior frontal opercularis.   

For anticipation (how baseline activity predicts value judgement) they found that OFC pre-stimulus activity was significantly associated with value signaling, but not the PHC activity. In the post-stimulus window, they found a significant association with both food and non-food item likeability rating for all regions, indicative of the generality of the signal. They also found that these regions responded to value in a distractive, or non-value task, reflective of automaticity in subjective valuation. Finally, they found that brain valuation system activity was also associated with a quadratic form of likeability ratings (quadraticity), indicative of a co-representation of confidence.

What's the impact?

This study identified a brain network important for valuation in decision making. Further, these findings indicate this network’s involvement in anticipation, generality, and automaticity in decision making. This work provides important evidence for how the brain assigns value to options during decision making and may help us understand the mechanisms of irrational judgement.

Alizee Lopez-Persem et al. Four core properties of the human brain valuation system demonstrated in intracranial signals. Nature Neuroscience (2020). Access the original scientific publication here