Mushroom bodies

Mushroom bodies are usually described as neuropils, i.e., as dense networks of neuronal processes (dendrite and axon terminals) and glia. They get their name from their roughly hemispherical calyx, a protuberance that is joined to the rest of the brain by a central nerve tract or peduncle.

Most of our current knowledge of mushroom bodies comes from studies of a few species of insect, especially the cockroach Periplaneta americana, the honey bee Apis mellifera,[7] the locust and the fruit fly Drosophila melanogaster. Studies of fruit fly mushroom bodies have been particularly important for understanding the genetic basis of mushroom body functioning, since their genome has been sequenced and a vast number of tools to manipulate their gene expression exist.

In the insect brain, the peduncles of the mushroom bodies extend through the midbrain. They are mainly composed of the long, densely packed nerve fibres of the Kenyon cells, the intrinsic neurons of the mushroom bodies. These cells have been found in the mushroom bodies of all species that have been investigated, though their number varies. Fruit flies, for example, have around 2,500, whereas cockroaches have about 200,000.

A locust brain dissection to expose the central brain and carry out electro-physiology recordings can be seen here. [8]

Historically, it was believed that only insects had mushroom bodies, because they were not present in crabs and lobsters. However, their discovery in the mantis shrimp in 2017 lead to the later conclusion[2] that the mushroom body is the ancestral state of all arthropods, and that this feature was later lost in crabs and lobsters.[2]

Mushroom bodies are best known for their role in olfactory associative learning. These olfactory signals are received from dopaminergic, octopaminergic, cholinergic, serotonergic, and GABAergic neurons outside the MB.[9] They are largest in the Hymenoptera, which are known to have particularly elaborate control over olfactory behaviours. However, since mushroom bodies are also found in anosmic primitive insects, their role is likely to extend beyond olfactory processing. Anatomical studies suggest a role in the processing of visual and mechanosensory input in some species.[10] In Hymenoptera in particular, subregions of the mushroom body neuropil are specialized to receive olfactory, visual, or both types of sensory input.[11] In Hymenoptera, olfactory input is layered in the calyx. In ants, several layers can be discriminated, corresponding to different clusters of glomeruli in the antennal lobes, perhaps for processing different classes of odors.[4][12] There are two main groups of projection neurons dividing the antennal lobe into two main regions, anterior and posterior. Projection neuron groups are segregated, innervating glomerular groups separately and sending axons by separate routes, either through the medial-antenno protocerebral tract (m-APT) or through the lateral-antenno protocerebral tract (l-APT), and connecting with two layers in the calyx of the mushroom bodies. In these layers the organization of the two efferent regions of the antennal lobe is represented topographically, establishing a coarse odotopic map of the antennal lobe in the region of the lip of the mushroom bodies.[4][12]

Mushroom bodies are known to be involved in learning and memory, particularly for smell, and thus are the subject of current intense research. In larger insects, studies suggest that mushroom bodies have other learning and memory functions, like associative memory, sensory filtering, motor control, and place memory. Research implies that mushroom bodies generally act as a sort of coincidence detector, integrating multi-modal inputs[4] and creating novel associations, thus suggesting their role in learning and memory.[13] Recent work also shows evidence for the involvement of the mushroom body in innate olfactory behaviors through interactions with the lateral horn,[14][15] possibly making use of the partially stereotyped sensory responses of the mushroom body output neurons (MBONs) across individuals.[16] Although the connections between the projection neurons and the Kenyon cells are random (i.e., not stereotyped across individuals),[17] the stereotypy in MBON responses is made possible by the dense convergence of many Kenyon cells onto a few MBONs along with other network properties.[16]

Information about odors may be encoded in the mushroom body by the identities of the responsive neurons as well as the timing of their spikes.[18] Experiments in locusts have shown that Kenyon cells have their activity synchronized to 20-Hz neural oscillations and are particularly responsive to projection neuron spikes at specific phases of the oscillatory cycle.[19]

Mushroom body lobes of D. melanogaster, with α/β, α’/β’, and γ neurons visible. From Davis, 2011[20]

We know that mushroom body structures are important for olfactory learning and memory in Drosophila because their ablation destroys this function.[21] The mushroom body is also able to combine information from the internal state of the body and the olfactory input to determine innate behavior.[22] The exact roles of the specific neurons making up the mushroom bodies are still unclear. However, these structures are studied extensively because much is known about their genetic make-up. There are three specific classes of neurons that make up the mushroom body lobes: α/β, α’/β’, and γ neurons, which all have distinct gene expression. A topic of current research is which of these substructures in the mushroom body are involved in each phase and process of learning and memory.[23] Drosophila mushroom bodies are also often used to study learning and memory and are manipulated due to their relatively discrete nature. Typically, olfactory learning assays consist of exposing flies to two odors separately; one is paired with electric shock pulses (the conditioned stimulus, or CS+), and the second is not (unconditioned stimulus, or US). After this training period, flies are placed in a T-maze with the two odors placed individually on either end of the horizontal ‘T’ arms. The percent of flies that avoid the CS+ is calculated, with high avoidance considered evidence of learning and memory.[24]

PKA dynamics

Spatial regulation of PKA dynamics in Drosophila mushroom body.

Protein kinase A (PKA) has been found to play an important role in learning and memory in Drosophila.[33] When calcium enters a cell and binds with calmodulin, it stimulates adenylate cyclase (AC), which is encoded by the rutabaga gene (rut).[34] This AC activation increases the concentration of cAMP, which activates PKA.[34] When dopamine, an aversive olfactory stimulant, is applied it activates PKA specifically in the vertical mushroom body lobes.[34] This spatial specificity is regulated by the dunce (dnc) PDE, a cAMP-specific phosphodiesterase. If the dunce gene is abolished, as found in the dnc mutant, the spatial specificity is not maintained. In contrast, an appetitive stimulation created by an octopamine application increases PKA in all lobes.[34] In the rut mutant, a genotype in which the rutabaga is abolished, the responses to both dopamine and octopamine were greatly reduced and close to experimental noise.

Acetylcholine, which represents the conditioned stimulus, leads to a strong increase in PKA activation compared to stimulation with dopamine or octopamine alone.[34] This reaction is abolished in rut mutants, which demonstrates that PKA is essential for sensory integration.[34] The specificity of activation of the alpha lobe in the presence of dopamine is maintained when dopamine is in combination with acetylcholine.[34] Essentially, during a conditioning paradigm when a conditioned stimulus is paired with an unconditioned stimulus, PKA exhibits heightened activation. This shows that PKA is required for conditioned learning in Drosophila melanogaster.

Stimulus → output responses are the product of pairs of excitation and inhibition. This is the same pattern of organisation as with mammals' brains. These patterns may, as with mammals, precede action. As of 2021 this is an area only recently elucidated by Zwaka et al 2018, Duer et al 2015, and Paffhausen et al 2020.[7]

• Cerebral cortex

• Standard BeeBrain