hypothalamus and food intake

No matter what technologies surround a person, his body remains a biological system (at least for now), which obeys a number of laws of biology. Some of these laws are physiological needs, namely water, food, sleep, sex, etc. When there is a shortage of one or another, the brain begins to actively send us signals. For example, if we are hungry, then we clearly need to eat. However, the process of eating food is much more complex from a cellular point of view than it might seem. Scientists from the University of Erlangen-Nuremberg (Nuremberg, Germany) conducted a study in which they found that signals run back and forth between neurons until a person replenishes the necessary amount of energy, thereby ensuring a balance of consumption (without overeating or undereating). A malfunction of this process leads to eating disorders (anorexia, overeating). What exactly did the scientists establish during their observations, and how to use the knowledge gained in practice? We will find the answers to these questions in the scientists' report.

Research basis

In humans and other mammals, innate behaviors maintain physiological homeostasis, organize interactions with conspecifics to ensure safety and reproduction of the species, and facilitate learning through exploratory responses to novelty. Innate behaviors are regulated by the hypothalamus, which processes a variety of chemical, sensory, and cognitive control signals. Hypothalamic output to the forebrain and midbrain is primarily transmitted by the anatomically complex and functionally diverse lateral hypothalamus (LH from

lateral hypothalamus

), which is critical for the coordination of multiple innate behaviors. Seminal studies have shown that LH lesions acutely reduce food intake and lead to starvation, whereas LH stimulation induces feeding urges. These effects are mediated by functionally and neurochemically distinct LH neurons that receive appetitive and anorexic signals from the hypothalamus, basal forebrain, and hindbrain, respond to glucose, and project to distinct regions in the feeding circuit.

Distinct populations of LH neurons participate in food intake and reward depending on their molecular identity and projections. Circuits mediating interactions with conspecific individuals also involve the LH, where cells with partially different molecular identities are excited or suppressed during social behavior. Interactions within and between these neuronal populations allow behavioral prioritization according to homeostatic needs and sensory cues.

Studies of another major hypothalamic output, the paraventricular nucleus, have raised the possibility that distinct subpopulations of cells with overlapping molecular profiles may be involved in different phases of feeding, including meal initiation, satiety enhancement, and meal termination. Orosensory and gastrointestinal stretch signals are processed by distinct populations of a key hypothalamic afferent area in the brainstem, generating consistent negative feedback signals during feeding. It is currently unknown whether there are feeding phase-specific cell populations in the hypothalamus and, importantly for LH, whether they are also involved in other innate behaviors.

Network oscillations influence the timing of neuronal discharge and coordinate signaling within and between brain regions, including the hypothalamus. Gamma oscillations have been implicated in feeding behavior. Food seeking in satiated mice involves coordinated gamma rhythmic activity in the LH, lateral septum, and prefrontal cortex, and is characterized by increased gamma-band coherence in the LH and ventral striatum upon repeated exposure to palatable food.

Gamma oscillations local field potentials* (LFP from local field potential) at the location of primary energy sensory neurons in the ventromedial hypothalamus were reduced in energy-deficient states in fasted mice and in mice with anorexia-induced weight loss during chronic exposure to the chemotherapeutic cisplatin.

Local Field Potentials (LFP)* — transient electrical signals generated in nerves and other tissues by the combined and synchronous electrical activity of individual cells (e.g. neurons) in that tissue.

The influence of gamma oscillations on neuronal activity in the hypothalamus has been demonstrated by studying single cycles of oscillations, where populations of cells preferentially fire during specific phases of the oscillation, thereby jointly determining efferent signaling. On the other hand, network synchronization during gamma oscillations involves increased activity in the population of participating neurons. The joint participation of neurons in oscillations, indicated by the degree of their correlated discharge, supports their direct and polysynaptic effects on each other and on their efferents, and influences synaptic plasticity. However, the organization of LH cell clusters according to their behavioral specificity during gamma oscillations, potentially supporting the coordination of functionally similar or different neurons, remains unknown.

In the work reviewed here, the researchers used spectral clustering to analyze the firing rate dynamics of LH neurons during different behaviors. Graphical spectral clustering, which outperforms traditional methods such as k-means and hierarchical clustering, is particularly useful for classifying the nonlinear dynamics of functionally heterogeneous neuronal populations. The researchers found that distinct populations of LH neurons recorded in the free-choice model were active at specific times during feeding episodes. Other populations of LH cells were more active during social interactions and exploration of novel objects than during feeding. LH gamma oscillations were associated with either increased or decreased co-activation of neurons compared to non-rhythmic epochs. LH cells from different groups showed distinct assembly patterns during slow and fast gamma oscillations, suggesting that different types of neural synchronization in the LH specifically support feeding and facilitate behavior-specific cell coordination.

Research results

Image #1

Firing of 1349 LH cells was recorded in four male mice using moving silicon probes in a free choice model during eating standard chow, interacting with a younger mouse of the same sex, or exploring a novel object.

These cells were recorded in parallel with at least one additional cell on the other stem of the silicon probe. Individual LH cells showed variable firing rate dynamics during the three behaviors. The scientists examined the multivariate firing rate vectors using an unsupervised approach, spectral clustering, which identifies communities of cells linked by a similarity graph (illustrated for the example subset of cells on 1a). To estimate the optimal number of clusters and k-NN (k-nearest neighbori.e. nearest neighbor) stability analysis was performed using ARI (adjusted Rand indexi.e. adjusted Rand index) for different parameter combinations, which gave high stability scores for the number of clusters from 2 to 11 with k-NN greater than 20 (1b).

Rand Index* (named after William M. Rand) in statistics, and particularly in data clustering, is a measure of the similarity between two clusters of data.

Mutual information* — a statistical function of two random variables that describes the amount of information contained in one random variable relative to the other.

Further evaluation of AMI distributions (

adjusted mutual information score

i.e. the adjusted estimate

mutual information*

) and ARI showed the most stable clustering using 2 and then 7 clusters (

). To compare the clustering quality between the 2- and 7-cluster approaches (using k-NN = 21), the scientists calculated the intra-cluster correlation ratio, which indicates the correlation of data within clusters compared to the correlation of data between clusters. The dataset segmented into 7 clusters showed higher intra-cluster correlations than the 2-cluster approach, as well as inter-cluster correlations (

–

). Thus, the segmentation of the data into 7 clusters was highly reproducible and correctly represented the similarity of firing rate dynamics between lateral hypothalamus (LH) neurons.

Image #2

To functionally characterize LH cells grouped based on similar firing rate dynamics, the scientists examined their activity in relation to innate behaviors. Their firing rate patterns were examined during feeding, exploration, and social interaction. Notably, although behaviors were not predetermined by label in the unsupervised classification, cells in different clusters (hereafter populations) fired in close correspondence to these behaviors (2a). The cumulative probability distributions of the times at which cells fired at their maximum rate during their preferred behavior (i.e., with the highest peak firing rate among the three behaviors) were not uniform across cells from the five populations active during feeding (2b). Four of these feeding-related populations clearly corresponded to different feeding phases over time (2a, 2b). These distinct feeding phase populations were sequentially active from feeding onset (FOn), through early and late feeding (EF and LF), to feeding termination (FOff). Cells from the exploration and social interaction populations (Exp and Sol) were uniformly activated during these respective behaviors and, together with a similar fifth feeding population (Fd), varied their average firing rate as a function of behavior.

It is important to note that populations were evident not only in the pooled data from multiple mice, but also within individual data (2c). Moreover, within each animal, the proportion of cells in each identified population was similar, with the exception of one animal with a lower number of neurons recorded (2c).

Feeding episodes were, on average, six times longer than exploration or social interaction episodes (2d). Thus, different feeding phase populations could be activated by a gradual increase in feeding duration or, alternatively, by signaling phases of feeding episodes, regardless of their duration. The duration of feeding episodes was similar among the different feeding-related populations (2e). Thus, these LH populations signal phases rather than durations of feeding.

Image #3

Differences in the timing of neural activity during feeding and in different behaviors are evident in individual cells (2a; Figure 3) were further elucidated by examining the dynamics of the average firing rate. During feeding, cells from the feeding populations showed consistently high levels of activity, signaling the relative timing of feeding episodes (4a, 4b; 2b). In contrast, during exploration and social interaction, cells from all populations showed transiently uniform activity, with cells from the exploration and social interaction populations firing at the highest rates during these behaviors (4a, 4b).

Image #4

The scientists then examined the activity of individual LH cells from each population during various behaviors. Activity during feeding was strongly inversely correlated with activity during exploration and social interaction for all populations (4c). However, activation during social interaction did not generally predict responding during exploration, suggesting an independent representation of these behaviors in the lateral hypothalamus (4c).

Image #5

The scientists next examined the anatomical localization of neurons from the recorded populations within the LH. For this analysis, 505 cells recorded from three mice in close anteroposterior planes (1.34-1.58 posterior bregma*; 5a).

Bregma* — an anthropometric point located on the roof of the skull and corresponding to the junction of the coronal (formed by the junction of the parietal bones and the frontal bone) and sagittal (formed by the junction of two parietal bones) sutures.

Two-dimensional probability density distributions of recorded cells were determined in mediolateral and dorsoventral locations, i.e., in accordance with the silicon probe rods and their progression, respectively. Probability density distributions were calculated for individual LH populations and for all populations together, the difference between the two probability densities at each location indicates location-specific enrichment of cells from a particular population (

). The FOn, EF and FOff populations showed correlated anatomical distributions, they were enriched in the lateral regions of the LH without clear dorsoventral differences (

). In contrast, the localization pattern of LF cells active during the late feeding phase was distinct from FOn, EF, Fd, and Exp and moderately correlated with FOff and Sol. LF cells were enriched in the dorsomedial and ventrolateral region of the LH (

). The localizations of Fd, Sol and Exp, uniformly active during individual innate behaviors, were distinct and poorly correlated with the localization of cells active during feeding phases.

The populations of LH feeding-phase neurons also had different electrophysiological properties. On the one hand, cells from feeding populations active during the second half of feeding had lower average firing rates throughout the recording than FOn and EF cells active at the beginning of feeding or cells from populations uniformly active throughout the behavior. On the other hand, the spike width of late feeding-phase LF cells was larger than that of early feeding-phase EF and individual populations of uniformly active cells.

In aggregate, populations active in the first half and later, but not until the end of feeding, may be enriched GABA* and orexin cells, respectively, two anatomically mixed types of LH cells.

GABA* (gamma Aminobutyric acid, i.e. gamma-aminobutyric acid or GABA) is an organic compound, a non-proteinogenic amino acid, the most important inhibitory neurotransmitter of the central nervous system (CNS) of humans and other mammals. Aminobutyric acid is a biogenic substance. It is contained in the CNS and takes part in neurotransmitter and metabolic processes in the brain.

In particular, the localization of LF cells at the analyzed anteroposterior level dorsal and lateral to the fornix and the physiological properties of LF correspond to those of orexin cells. Lateral areas of LH, enriched in the Sol population, are characterized by MCH neurons (

melanin-concentrating hormone

i.e., a hormone that concentrates melanin), involved in social behavior.

Image #6

Optogenetic studies show that populations of hypothalamic neurons strongly influence innate behavior when activated synchronously. During spontaneous behavior, such collective activity is thought to require sufficiently strong reciprocal and/or afferent coupling of behavior-specific cells to ensure their temporal coordination. Formation of cell clusters may be facilitated by gamma oscillations, as suggested by hippocampal studies.

To investigate clusters of cells from functional LH populations, the scientists first characterized co-firing across the entire population of recorded LH neurons during gamma oscillations and epochs without network oscillations (hereafter referred to as non-rhythmic epochs; 6a, 6b). CCGs (from cross-correlogrami.e., cross-correlograms) of 10,838 pairs of LH neurons recorded with different silicon probe rods, using spikes firing during locally recorded gamma oscillations or non-rhythmic epochs as trigger events. The preferential co-firing during non-rhythmic epochs compared to gamma oscillation epochs was then quantified as the co-firing ratio (CFR co-firing ratio) during these epochs for each pair of cells (6c).

Most cell pairs increased co-firing during gamma oscillations compared to non-rhythmic epochs, as indicated by CFR < 1 (6d–6g). This is consistent with the general increase in synchronization during gamma oscillations predicted by computational models and demonstrated in the hippocampus and in the lateral septum, the main forebrain input to the LH. At the same time, a surprisingly high proportion (~40%) of LH neurons reduced their cooperative activity to varying degrees compared to nonrhythmic epochs (6d–6g). Slow and fast gamma oscillations were associated with different changes in co-activation in individual cell pairs. Specifically, cell pairs with low CFR during slow gamma oscillations (i.e., they were more coordinated during slow gamma oscillations) showed higher CFR during fast gamma oscillations (i.e., lower coordination during fast gamma oscillations; data points below the diagonal in 6h). Conversely, many cell pairs with high CFR during slow gamma oscillations showed lower CFR during fast gamma oscillations (data points above the diagonal in 6h).

Image #7

Finally, the scientists investigated whether the increase in co-activation during gamma oscillations relative to non-rhythmic epochs depends on the behavioral identity of the neurons. To this end, the scientists compared the CFRs calculated for CCG cells from the same or different LH functional populations according to their possible behavioral hierarchy using the original and shifted spike trains (graph above).

Cells from feeding-related populations, but not from populations predominantly active during different behaviors, were more coordinated during slow gamma oscillations than during nonrhythmic epochs. Accordingly, cell coordination in some population groups (EF and Sol, FOff and Exp) was even reduced during slow gamma oscillations. A notable exception was the increased coordination of LF and Sol cells during both slow and fast gamma oscillations (graph below).

Image #8

In contrast, during fast gamma oscillations, coordination between multiple neuronal populations was enhanced. This included not only greater co-activation of cells between and within feeding-related populations during fast gamma, but also greater co-activation of neurons from populations active during feeding, social interaction, and exploration of novel objects (Sol and LF, Sol and FOff, Exp and EF, Exp and Fd). Consistent with the uncorrelated firing rates of individual cells between social and exploratory behavior (4c), co-activation of neurons from the Sol and Exp populations did not differ during slow or fast gamma oscillations and non-rhythmic epochs. Thus, the high degree of specialization of LH populations in feeding control is accompanied by a distinct mode of network synchronization of slow gamma oscillations that predominantly promotes interactions of LH neurons within and between feeding-related populations.

For a more detailed look at the nuances of the study, I recommend taking a look at scientists' report.

Epilogue

In the work we reviewed today, scientists decided to study the details of what neural activity occurs during eating, namely during the different stages of this behavior. The main conclusion of the study is that neural activity jumps from one neural population to another until the body receives the necessary amount of energy, that is, until the person is full. Disruption of this activity can lead to undereating or overeating.

The human body has a number of needs that directly play a role in its survival, including food. When we eat, we replenish our energy, and the hypothalamus regulates this process. Being a kind of computing computer, the hypothalamus collects and processes information from the environment and from our own body (for example, the perception of the time of day or blood sugar levels). As a result of analyzing the data received, the hypothalamus triggers innate behavior. In the case of the time of day, for example, this is the desire to sleep in the dark, and in the case of hunger – the search for food.

Scientists note that during food consumption, a person quickly switches from “appetizing” to “consumer” behavior. In other words, hunger can subside quite quickly, but this does not mean that the body has replenished the required amount of energy, which is why the “consumer” phase of nutrition is activated.

Scientists analyzed the electrical activity of a specific area of the hypothalamus to determine which neurons are activated at any given time during a meal. As a result, four groups of neurons (populations) were identified that are consistently activated during a meal. Each of these groups processes specific information (blood sugar level, the number of hunger hormones, the stretching of the stomach walls (feeling of fullness), etc.). That is, the fourth group of neurons can transmit more information about the state of the stomach than, for example, the first group. This allows the hypothalamus to control the process of eating, ensuring that there is no overeating or undereating.

Another important aspect of the study was the study of neuron interactions not only between populations, but also within populations. Scientists have found that such interactions can occur extremely frequently (several dozen times per second) and require oscillations at the same frequency. In other words, neurons involved in feeding behavior “communicate on the same wavelength” within and between populations. The most interesting thing is that neurons responsible for other behavior (learning new things or social interaction) communicate at a different frequency, thereby not creating barriers for “feeding” neurons.

With the knowledge gained, as well as the fact that the frequency of neurons can be influenced from the outside (for example, using oscillating magnetic fields), it is possible to develop a method for improving communication between nutrition neurons if it has been disrupted due to injury or illness. In the future, this will allow the process of food consumption to be controlled at the neural level and human behavior to be corrected, thereby reducing or completely eliminating malnutrition or overeating.

A bit of advertising

Thank you for staying with us. Do you like our articles? Want to see more interesting materials? Support us by placing an order or recommending to your friends,

Cloud VPS for developers from $4.99

a unique analogue of entry-level servers, which we have created for you:The whole truth about VPS (KVM) E5-2697 v3 (6 Cores) 10GB DDR4 480GB SSD 1Gbps from $19 or how to properly share a server?

(RAID1 and RAID10 options available, up to 24 cores and up to 40GB DDR4).

Dell R730xd 2x cheaper in Maincubes Tier IV data center in Amsterdam? Only with us 2 x Intel TetraDeca-Core Xeon 2x E5-2697v3 2.6GHz 14C 64GB DDR4 4x960GB SSD 1Gbps 100 TV from $199 in the Netherlands! Dell R420 – 2x E5-2430 2.2Ghz 6C 128GB DDR3 2x960GB SSD 1Gbps 100TB – from $99! Read about How to build a corporate class infrastructure using Dell R730xd E5-2650 v4 servers costing 9000 euros for pennies?