 | |  | |  | |  | |  | |  | |  | |  | |  |
 | |  | |  | |  | |  | |  | |  | |  | |  | |  |
| The Insect Brain |
The basic insect nervous system bauplan consists of a series of body segments, each equipped with a pair of connected
ganglia, with a paired nerve cord connecting adjacent ganglia in each segment. The ganglia are bulbous structures
consisting of neuron cell-bodies and supporting or glial cells and acts as a local processor or computer. the ganglia are
interconnected by neurons, constituing a computer network. This plan is variously modified in the various types, but in all
cases the ganglia of the head segment form a fused mass, situated above the oesophagus (esophagus) of the gut, and
called the supraoesophageal ganglion. This is connected by a pair of nerve trunks (connectives or commissures) that
course around the oesophagus on either side and join to the suboesophageal ganglion (SOG or SEG) situated beneath
the oesophagus. Many consider only the supraoesophageal ganglion to constitute the insect brain, others (including myself)
consider the SOG as part of the brain.
ganglia, with a paired nerve cord connecting adjacent ganglia in each segment. The ganglia are bulbous structures
consisting of neuron cell-bodies and supporting or glial cells and acts as a local processor or computer. the ganglia are
interconnected by neurons, constituing a computer network. This plan is variously modified in the various types, but in all
cases the ganglia of the head segment form a fused mass, situated above the oesophagus (esophagus) of the gut, and
called the supraoesophageal ganglion. This is connected by a pair of nerve trunks (connectives or commissures) that
course around the oesophagus on either side and join to the suboesophageal ganglion (SOG or SEG) situated beneath
the oesophagus. Many consider only the supraoesophageal ganglion to constitute the insect brain, others (including myself)
consider the SOG as part of the brain.
The relative size of the brain
species. That of the diving
beetle Dytiscus is about 1/400
of the total body size, that of the
ant (Formica) is about 1/280 ,
and that of the bee about 1/174.
The brain is generally larger in
those insects that have more
complex social lives.
Although much smaller than a
human brain, containing only
one thousandth as many cells, it
is still immensely complex.
There is also less replication of
function - fewer neurons
perform each function.
species. That of the diving
beetle Dytiscus is about 1/400
of the total body size, that of the
ant (Formica) is about 1/280 ,
and that of the bee about 1/174.
The brain is generally larger in
those insects that have more
complex social lives.
Although much smaller than a
human brain, containing only
one thousandth as many cells, it
is still immensely complex.
There is also less replication of
function - fewer neurons
perform each function.
The supraoesophageal ganglion consists of several fused ganglia or lobes. The paired ganglia of the first (frontmost) head
segment form the protocerebrum, concerned with vision, time-keeping, higher functions, memory and combining
information from different sensory modalities. Those of the segment segment form the deutocerebrum, which is concerned
with processing sensory inputs from the antennae, and also the labial palps and parts of the tegument (body wall). the
ganglia of the third segment form the relatively small tritocerebrum.
segment form the protocerebrum, concerned with vision, time-keeping, higher functions, memory and combining
information from different sensory modalities. Those of the segment segment form the deutocerebrum, which is concerned
with processing sensory inputs from the antennae, and also the labial palps and parts of the tegument (body wall). the
ganglia of the third segment form the relatively small tritocerebrum.
AL, antennal lobe; DV, dorsal blood vessel; L, lamina; LCB, lower central body; Lo,
lobula; M, medulla; MB, mushroom body; PB, procerebral bridge; PI, pars
intercerebralis; T, tritocerebrum; UCB, upper central body; XL: accessory lobe.
Protocerebrum
The optic lobes of the fly (an insect with particularly good vision) contains about
76% of the brain's neurons. The optic lobe connects directly to the sensory cells
(retinula cells) in the retina of the compound eye. It contains three distinct regions
(neuropils): the lamina, medulla and lobula, where processing of visual signals
begins. The protocerebrum also receives inputs via the ocelli, when present, via
the ocellar nerves.
The mushroom bodies (MB, corpora pedunculata, 'stalked bodies') are best
developed in social insects, making up 20% of the brain of the bee and 50% of the
brain of worker ants (Formica). These are thought to function as higher centres
responsible for the most sophisticated computations occurring in the insect brain.
Each consists of a topmost cap and a stalk or peduncle (which branches into at
least two lobes). The cap consists of a pair of cup-like structures, the medial calyx
and the lateral calyx (plural of calyx is calyces). The mushroom bodies receive
sensory inputs from the lobula of the optic lobe and from the antennal lobes of the
deutocerebrum. Most sensory inputs enter the MB through the calyx. There are
about 1000 to 100 000 specialised neurons, called Kenyon cells, in each
mushroom body. These neurons have tree-like branching dendrites which receive
inputs in the calyces of the MB, a single axon which extends down the stalk of the
MB and then gives of branches to two lobes of the MB. Dragonfly mushroom
bodies have no calyces and no Kenyon cells. The mushroom bodies are also
involved in learning, and in the honeybee have been shown to process memories,
transferring data from short-term memory (STM) into long-term memory (LTM).
The central body receives inputs from the mushroom bodies and integrates
sensory inputs from different sensory modalities (such as small and vision) -
so-called multimodal sensory perception. It functions as an activating centre,
switching on appropriate locomotor activity patterns which are central programs
located in the thoracic ganglia. That is it instructs the thoracic ganglia which
programs to run - programs that control the legs and wings. These hard-wired
programs are sometimes called central pattern generators and require no
sensory input for their execution, though sensory inputs may start and stop these
programs or modify them slightly.
The pars intercerebralis is a mass of cell bodies, including neurosecretory
cells which send their axons to the pair of corpora cardiaca (see the
neuroendocrine system in insect development). The corpora cardiac are
sometimes fused into a single medial ganglion. They send out nerves to innervate
the dorsal blood vessel, forming a cardio-aortic system, which controls the rate
of heart beat, as well as having a secretory hormona lfunction.
lobula; M, medulla; MB, mushroom body; PB, procerebral bridge; PI, pars
intercerebralis; T, tritocerebrum; UCB, upper central body; XL: accessory lobe.
Protocerebrum
The optic lobes of the fly (an insect with particularly good vision) contains about
76% of the brain's neurons. The optic lobe connects directly to the sensory cells
(retinula cells) in the retina of the compound eye. It contains three distinct regions
(neuropils): the lamina, medulla and lobula, where processing of visual signals
begins. The protocerebrum also receives inputs via the ocelli, when present, via
the ocellar nerves.
The mushroom bodies (MB, corpora pedunculata, 'stalked bodies') are best
developed in social insects, making up 20% of the brain of the bee and 50% of the
brain of worker ants (Formica). These are thought to function as higher centres
responsible for the most sophisticated computations occurring in the insect brain.
Each consists of a topmost cap and a stalk or peduncle (which branches into at
least two lobes). The cap consists of a pair of cup-like structures, the medial calyx
and the lateral calyx (plural of calyx is calyces). The mushroom bodies receive
sensory inputs from the lobula of the optic lobe and from the antennal lobes of the
deutocerebrum. Most sensory inputs enter the MB through the calyx. There are
about 1000 to 100 000 specialised neurons, called Kenyon cells, in each
mushroom body. These neurons have tree-like branching dendrites which receive
inputs in the calyces of the MB, a single axon which extends down the stalk of the
MB and then gives of branches to two lobes of the MB. Dragonfly mushroom
bodies have no calyces and no Kenyon cells. The mushroom bodies are also
involved in learning, and in the honeybee have been shown to process memories,
transferring data from short-term memory (STM) into long-term memory (LTM).
The central body receives inputs from the mushroom bodies and integrates
sensory inputs from different sensory modalities (such as small and vision) -
so-called multimodal sensory perception. It functions as an activating centre,
switching on appropriate locomotor activity patterns which are central programs
located in the thoracic ganglia. That is it instructs the thoracic ganglia which
programs to run - programs that control the legs and wings. These hard-wired
programs are sometimes called central pattern generators and require no
sensory input for their execution, though sensory inputs may start and stop these
programs or modify them slightly.
The pars intercerebralis is a mass of cell bodies, including neurosecretory
cells which send their axons to the pair of corpora cardiaca (see the
neuroendocrine system in insect development). The corpora cardiac are
sometimes fused into a single medial ganglion. They send out nerves to innervate
the dorsal blood vessel, forming a cardio-aortic system, which controls the rate
of heart beat, as well as having a secretory hormona lfunction.
| Did you know? The insect brain contains about 100 000 to 1000 000 neurons, compared to one billion (one thousand million) found in the human brain. |
Biological Clocks
Another function associated with the protocerebrum is time-keeping. Insect activity is timed with the daily light/dark cycle - the
circadian cycle ('ciracdian' means 'about a day', the exact time being set each day according to environmental cues such
as the length of daylight). This timing is due to internal clocks within the insect, which update themselves according to
external cues from the environment (zeitgeibers or time-givers) such as the number of hours of light and dark. (This
resetting by use of external signals enables the insect to adjust to different local conditions depending, for example, on
latitude). Many body parts and organs have their own circadian clocks, indeed each cell appears capable of keeping time,
but these appear to be set and synchronised by a central master clock, which resides in the protocerebrum and is both
neural and hormonal. In some insects, a master clock is found in each optic lobe, which makes sense as these process light
signals. There is also a daily movement of screening pigments in the ommatidia of the compound eye, as the insect adjusts
to night-time darkness by increasing the sensitivity of its retina (it will continue to do this at the correct time for days when
kept in constant light or dark for several days, so the response is coordinated, in part, by a central clock). Severing of the
optic lobes prevents these clocks from synchronising bodily activities. In other species, however, the clock is only abolished if
the brain is cut in two, which suggests that it may reside in the central body.
Deutocerebrum
This consists of two nerve centres - the main antennal lobe (AL) and the smaller antennal mechanosensory and motor
centre (AMMC) or dorsal lobe. The AL receives inputs from the third (terminal) antennal segment (the flagellum, which is
made-up of sub-segments called flagellomeres) via the antennal nerves. It contains from less than 10 to more than 200
sub-centres called glomeruli (singular glomerulus). Inputs to the AL appear to be mainly or exclusively from chemoreceptors
(i.e. chemical sensors - olfactory and gustatory, smell and taste) on the flagellum. Each antenna sends signals to the AL on
the same side of the head (ipsilateral pathways) although some may also send signals to the AL on the opposite side
(contralateral pathways).
Each glomerulus is a region of neuropil (nerve cell processes and synapses) where computations occur. It is thought that
each glomerulus may, in some species at least, receive inputs from a specific class of receptor (sensor) on the antenna. For
example, in the males of some species there is a specially large glomerulus, called the macroglomerular complex (MGC)
which receives inputs from pheromone olfactory sensors on the antenna.
The AL does not receive one input line from each chemoreceptor, as sensors of the same type converge - their axons fuse
into a smaller number of axons in the antennal nerve (typically inputs from 15 sensors are combined, a 15:1 ratio). These
sensory input axons, and also input axons from the CB of the protocerebrum, synapse with local interneurones within the AL
(amacrine cells). Outputs from the AL are carried along the axons of output neurons to the MB of the protocerebrum.
The AMMC receives mechanosensory inputs from mechanosensors (mechanoreceptors)on the first two antennal segments
(scape and pedicel) via the antennal nerves. It also sends motor outputs to the muscles of the scape. It also receives inputs
from mechanosensors on the labial palps, some tegument (body wall) mechanosensors, and some inputs from the flagellum
(possibly from the mechanosensors found on the flagellum). The antennal nerve is therefore a mixed nerve - containing both
sensory and motor axons. Some of the antennal mechanoreceptors also send outputs to the SOG, the protocerebrum and
the thoracic ganglia.
Tritocerebrum and Stomatogastric System
The frontal ganglion (FG) is an additional free and single (unpaired) median (median = in the midline) ganglion that is
connected by a pair of bilateral connectives to the tritocerebrum. A single medial recurrent nerve runs back up to a ganglion
situated beneath and behind the supraoesophageal ganglion. This ganglion may be called the stomachic ganglion or the
hypocerebral ganglion (HG). In the locust, the HG sends out one pair of outer oesophageal nerves (and one pair of inner
oesophageal nerves (ventricular nerves). Each of the latter terminates in a ventricular ganglion (ingluvial ganglion) on the
crop of the foregut (see insect nutrition). These then control crop movements. In Dytiscus, it has been shown that the FG
also controls swallowing. Thus, the tritocerebrum and frontal ganglion control the foregut, forming the stomatogastric
system. The tritocerebrum also innervates the labrum.
Suboesophageal Ganglion
The suboesophageal ganglion (SOG) and the segmental ganglia of the double ventral nerve-cord each send out pairs of
nerves, one of which innervates the pair of spiracles on that segment and so help regulate breathing. (In some insects the
segmental ganglia are absent, e.g. in Dytiscus, in which case the lateral abdominal nerves send out nerves to innervate the
spiracles). The SOG is a composite ganglion, formed by fusion of the ganglia from the mandibular, maxillary and labial
segments of the head and the SOG also sends out nerves to the mouthparts (mandibles, palps, etc.) and so controls feeding
behaviours.
The Ventral Nerve Cord
From the suboesophageal ganglion two connectives or nerve cords run back along the ventral side (underside) of the
insect. These connect to the thoracic ganglion of the first thoracic segment, T1, which is actually a pair of ganglia,
more-or-less fused into a single structure. T1 then gives off two connectives to the second thoracic ganglion, T2 and the
sequence continues with a chain of connected ganglia running throughout the length of the insect, in the basic plan. Thus,
we say that insects have a double ganglionated ventral nerve cord (VNC). Each ganglion functions as a local processor,
regulating the functions of its body segment. The thoracic ganglia are especially well-developed as they have to carry out
complex computations to generate patterns of movement in the legs and wings. These output patterns or central
programs are contained in the ganglia, but the brain is normally required to switch them on and off. Sensory inputs have
little effect on the basic patterns, but do modify them. For example, stress sensors in the wings feedback information to allow
fine-adjustments to the wings and control of the angle of attack and wing-twisting. Typically, however, the basic pattern of
movement is pre-coded.
Another function associated with the protocerebrum is time-keeping. Insect activity is timed with the daily light/dark cycle - the
circadian cycle ('ciracdian' means 'about a day', the exact time being set each day according to environmental cues such
as the length of daylight). This timing is due to internal clocks within the insect, which update themselves according to
external cues from the environment (zeitgeibers or time-givers) such as the number of hours of light and dark. (This
resetting by use of external signals enables the insect to adjust to different local conditions depending, for example, on
latitude). Many body parts and organs have their own circadian clocks, indeed each cell appears capable of keeping time,
but these appear to be set and synchronised by a central master clock, which resides in the protocerebrum and is both
neural and hormonal. In some insects, a master clock is found in each optic lobe, which makes sense as these process light
signals. There is also a daily movement of screening pigments in the ommatidia of the compound eye, as the insect adjusts
to night-time darkness by increasing the sensitivity of its retina (it will continue to do this at the correct time for days when
kept in constant light or dark for several days, so the response is coordinated, in part, by a central clock). Severing of the
optic lobes prevents these clocks from synchronising bodily activities. In other species, however, the clock is only abolished if
the brain is cut in two, which suggests that it may reside in the central body.
Deutocerebrum
This consists of two nerve centres - the main antennal lobe (AL) and the smaller antennal mechanosensory and motor
centre (AMMC) or dorsal lobe. The AL receives inputs from the third (terminal) antennal segment (the flagellum, which is
made-up of sub-segments called flagellomeres) via the antennal nerves. It contains from less than 10 to more than 200
sub-centres called glomeruli (singular glomerulus). Inputs to the AL appear to be mainly or exclusively from chemoreceptors
(i.e. chemical sensors - olfactory and gustatory, smell and taste) on the flagellum. Each antenna sends signals to the AL on
the same side of the head (ipsilateral pathways) although some may also send signals to the AL on the opposite side
(contralateral pathways).
Each glomerulus is a region of neuropil (nerve cell processes and synapses) where computations occur. It is thought that
each glomerulus may, in some species at least, receive inputs from a specific class of receptor (sensor) on the antenna. For
example, in the males of some species there is a specially large glomerulus, called the macroglomerular complex (MGC)
which receives inputs from pheromone olfactory sensors on the antenna.
The AL does not receive one input line from each chemoreceptor, as sensors of the same type converge - their axons fuse
into a smaller number of axons in the antennal nerve (typically inputs from 15 sensors are combined, a 15:1 ratio). These
sensory input axons, and also input axons from the CB of the protocerebrum, synapse with local interneurones within the AL
(amacrine cells). Outputs from the AL are carried along the axons of output neurons to the MB of the protocerebrum.
The AMMC receives mechanosensory inputs from mechanosensors (mechanoreceptors)on the first two antennal segments
(scape and pedicel) via the antennal nerves. It also sends motor outputs to the muscles of the scape. It also receives inputs
from mechanosensors on the labial palps, some tegument (body wall) mechanosensors, and some inputs from the flagellum
(possibly from the mechanosensors found on the flagellum). The antennal nerve is therefore a mixed nerve - containing both
sensory and motor axons. Some of the antennal mechanoreceptors also send outputs to the SOG, the protocerebrum and
the thoracic ganglia.
Tritocerebrum and Stomatogastric System
The frontal ganglion (FG) is an additional free and single (unpaired) median (median = in the midline) ganglion that is
connected by a pair of bilateral connectives to the tritocerebrum. A single medial recurrent nerve runs back up to a ganglion
situated beneath and behind the supraoesophageal ganglion. This ganglion may be called the stomachic ganglion or the
hypocerebral ganglion (HG). In the locust, the HG sends out one pair of outer oesophageal nerves (and one pair of inner
oesophageal nerves (ventricular nerves). Each of the latter terminates in a ventricular ganglion (ingluvial ganglion) on the
crop of the foregut (see insect nutrition). These then control crop movements. In Dytiscus, it has been shown that the FG
also controls swallowing. Thus, the tritocerebrum and frontal ganglion control the foregut, forming the stomatogastric
system. The tritocerebrum also innervates the labrum.
Suboesophageal Ganglion
The suboesophageal ganglion (SOG) and the segmental ganglia of the double ventral nerve-cord each send out pairs of
nerves, one of which innervates the pair of spiracles on that segment and so help regulate breathing. (In some insects the
segmental ganglia are absent, e.g. in Dytiscus, in which case the lateral abdominal nerves send out nerves to innervate the
spiracles). The SOG is a composite ganglion, formed by fusion of the ganglia from the mandibular, maxillary and labial
segments of the head and the SOG also sends out nerves to the mouthparts (mandibles, palps, etc.) and so controls feeding
behaviours.
The Ventral Nerve Cord
From the suboesophageal ganglion two connectives or nerve cords run back along the ventral side (underside) of the
insect. These connect to the thoracic ganglion of the first thoracic segment, T1, which is actually a pair of ganglia,
more-or-less fused into a single structure. T1 then gives off two connectives to the second thoracic ganglion, T2 and the
sequence continues with a chain of connected ganglia running throughout the length of the insect, in the basic plan. Thus,
we say that insects have a double ganglionated ventral nerve cord (VNC). Each ganglion functions as a local processor,
regulating the functions of its body segment. The thoracic ganglia are especially well-developed as they have to carry out
complex computations to generate patterns of movement in the legs and wings. These output patterns or central
programs are contained in the ganglia, but the brain is normally required to switch them on and off. Sensory inputs have
little effect on the basic patterns, but do modify them. For example, stress sensors in the wings feedback information to allow
fine-adjustments to the wings and control of the angle of attack and wing-twisting. Typically, however, the basic pattern of
movement is pre-coded.
Learning, Memory and Intelligence in Insects
Although the mushroom bodies of the brain have been shown to be involved in learning, ganglia other than the brain are
also capable of learning. Learning has been demonstrated in decapitated cockroaches! If a headless cockroach is wired so
that one of its legs receives an electric shock when lowered, then it will learn to avoid the shocks by keeping the leg raised!
(This is a classic experiment). The thoracic ganglia are responsible for controlling leg movement, and it is these ganglia that
learn the new behaviour. Intact cockroaches have a preference for darkness if given a choice between the illuminated half
of a chamber and the darkened half. However, if they receive electric shocks in the darkened half, then they will learn to
avoid the dark-half and remain in the light.
Habituation. An isolated cockroach leg exhibits another phenomenon related to learning - sensory habituation. If a
touch-sensitive bristle is stimulated on an isolated leg and the activity in the leg nerve recorded, then it will be found that the
strength of the stimulus diminishes if the stimulus is repeated rapidly, or sustained. This is due to fatigue in the periphery
nerve and sensor, and this ensures that insects respond to changes in the environment, and learn to ignore persistent
stimuli that are of no relevance. For example, body lice prefer rough fabrics, like wool, to smooth fabrics like silk. When
crossing from wool to silk they will keep turning as they try to find their way back onto the wool. However, if they fail to find
the wool (say if it is removed) then this behaviour stops as the lice learn to make do - they habituate. Habituated insects are
still responsive to changes in the stimulus, however. Thus, if the texture of the fabric for our louse on silk changes again,
then we expect it will respond again, and habituate again if necessary.
Learning one's way about - route learning. Some insects, such as ants and cockroaches, are capable of learning the route
of a maze, if the maze exit leads to reward, or if escaping from the maze avoids punishment. A cockroach will navigate a
maze to it home-pot, so long as the home-pot contains recognisable cockroach odours. On subsequent trials it will reach its
home-pot with increasing ease as it gradually learns the route. When compared to rats, ants learn mazes at half the speed:
a rat will master a maze after about 15 runs, an ant after about 30, though the ants still make a few errors. (Not bad when
you consider how tiny the ant brain is!). However, if a change is made to the maze, such as reversing the pattern, then rats
learn the 'new' maze more rapidly than before, recognising the similarity in the patterns and transferring their previous
learning to the new situation (transfer learning). In contrast, the ant starts all-over again, treating the maze as entirely
new. Thus ants have little or no capacity for transfer learning, that is they seem unable to apply what they have learnt to a
novel situation.
One of the most impressive feats of insect learning is locality learning. In addition to learning routes, insects can
recognise the locality in which their nest is situated, or in which food is found. This involves exploratory learning- the insect
will typically fly around a bit after leaving the nest, learning the position of many landmarks very rapidly and then leave. This
is latent learning, meaning that a period of time elapses between learning and reward. The reward occurs when the insect
returns home and locates its nest. This can be demonstrated by experiment. For example, in classic experiments on the
beewolf, Philanthus triangulum a type of digger wasp, which brings back food to its developing young in the nest, the wasp
will learn to recognise a ring of pine cones placed around the nest entrance and that the nest hole is the centre of this ring.
If the pine cones are displaced a few centimetres when the wasp leaves, however, it will return to the centre of the ring of
pine cones (only its nest is no longer there!).Generally, however, these insects are only temporarily confused by changes to
one or two landmarks, falling back on other landmarks further from the nest (and perhaps other cues like smell?).
Associative learning occurs when a stimulus, irrelevant by itself, is made relevant by pairing it with something meaningful,
like a food reward and the animal learns to identify the hitherto unconnected stimulus with food. For example, a cockroach
can be rewarded by being given food when presented with a particular, but unrelated, odour. It will then learn to associate
that odour with food and be attracted to it when foraging. Bees can be trained to associate a particular colour or pattern
9though their perception of shape is limited) with food.
Short-term and long-term memories. When an insect has just completed a task, learning is abolished if a new task is
undertaken immediately after. Learning requires a latent period of rest in-between activities. To some extent the same is
true of humans - learning becomes greatly enhanced if breaks are taken every 20 minutes or so during study. It is during
these rest periods that the brain processes the information (often subconsciously in humans) and the appropriate neural
pathways are reinforced. [Dreams in humans are especially curious - if activities are undertaken straight after waking then
dreams are often very easily forgotten, but if a few minutes are spent reflecting on them they may be remembered more
easily]. This can be explained by the existence of short-term memory (STM) and long-term memory (LTM). The nervous
system requires time to process information in short-term memory, making sense of it and discarding information deemed
irrelevant and transferring more useful information into long-term memory. In insects LTM typically retains information for
several days.
Do Insects Sleep? Hard to say. Insects certainly exhibit circadian cycles. Insects enter a dormant or semi-dormant state
when chilled. This may occur seasonally, during winter, or daily at nightfall. At nighttime insects may assume a particular
posture in which to rest, bees may hold onto vegetation with their tightly clamped jaws, for example. They may also return to
the same resting spot each night. During these quiescent states the body temperature drops and energy is conserved.
Interestingly, learning increases when insects are given rest between tasks or training sessions, as already mentioned, but
also if they are chilled during these rest periods. Could it be that when chilled into torpor at nighttime insects are forming
memories? It has been suggested that two chief functions of sleep are: energy conservation and memory formation. In this
case the insect nighttime torpor is not so very different. Many insects are, of course, nocturnal. Cockroaches are nocturnal
and learn better at night than during the day. It has been suggested that this encourages them to remember useful things
like lessons learned during nighttime foraging.
Temporal learning occurs in insects. As we have seen insects can measure time, by the use of internal body-clocks, and
they can also compensate for the changing position of the Sun in the sky when they use the Sun to navigate. To navigate
successfully by the Sun they must know what time it is, in the sense of what portion of daylight has passed. Observations
have shown that this may require learning, with young insects making mistakes by assuming that the Sun stays fixed in the
sky.
Insight learning is a higher form of learning, similar to transfer learning, in that it takes prior learning and applies that to a
new problem. One example is tool use. Some insects use tools, for example, some digger wasps fill in their nest burrow once
the young is mature and ready to pupate (and so requires protection but no more food). The female may then hold a small
'pebble' in her mandibles and use it to pound-down the earth, and then discard the pebble. However, this behaviour is not
intelligent in the sense that the insect reasoned a solution, rather it is inherited genetically and is instinctive and so not true
insight learning.
The Peripheral Nervous System
Peripheral nerves may be sensory or motory, but in insects are generally mixed. For example, the antennal nerve carries
many sensory fibres conveying inputs from the many antennal sensors to the brain, but it also contains some motor fibres
carrying output signals to the muscles in the base of the antenna. Animal nerve fibres, which are usually the axons of
neurons, are typically wrapped by insulating glial cells. In nonmyelinated axons, the wrapping may be a simple sheath
that loosely invests a group of axons. One function of this sheath is to ensure that the axons are bathed in a suitable salt
solution necessary for them to conduct impulses. In myelinated axons, however, each individual axon is tightly wrapped in its
own insulating sheath of material called myelin, which folds tightly around the axon several times, except at exposed
regions (such as the nodes of Ranvier in mammals, although these may be simple pores in other animals). This more
advanced type of wrapping is insulating and serves to speed-up nerve transmission. Students of vertebrate zoology often
mistakenly believe that only vertebrates possess myelinated axons (in addition to unmyelinated axons which also occur in
vertebrates). Myelinated axons occur in many invertebrates. Indeed, the fastest nerve transmission in the animal kingdom is
seen in the myelinated axons of certain shrimps, which conduct signals at about 200 m/s. Insects have an intermediate form
of insulation, which is like the myelin sheath, except that the myelin is wound loosely, leaving fluid-filled spaces between the
layers. Such a nerve fibre is intermediate between unmyelinated and myelinated and is called a tunicated nerve fibre.
Although Wigglesworth reported seeing the occasional classical myelinated nerve fibre in some insects, they are generally
considered absent from insects. Animals generally invest in the kind of sheathing that is adequate for their purposes. Large
and fast-moving animals, like humans, and small but very fast-moving animals like certain shrimps use myelin where
needed. insects are perhaps too small to require much, if any, use of myelin. The intermediate tunic seems quite sufficient
for their purposes. [A good external link to invertebrate myelin is: http://www5.pbrc.hawaii.edu/~danh/InvertebrateMyelin/ ].
Although the mushroom bodies of the brain have been shown to be involved in learning, ganglia other than the brain are
also capable of learning. Learning has been demonstrated in decapitated cockroaches! If a headless cockroach is wired so
that one of its legs receives an electric shock when lowered, then it will learn to avoid the shocks by keeping the leg raised!
(This is a classic experiment). The thoracic ganglia are responsible for controlling leg movement, and it is these ganglia that
learn the new behaviour. Intact cockroaches have a preference for darkness if given a choice between the illuminated half
of a chamber and the darkened half. However, if they receive electric shocks in the darkened half, then they will learn to
avoid the dark-half and remain in the light.
Habituation. An isolated cockroach leg exhibits another phenomenon related to learning - sensory habituation. If a
touch-sensitive bristle is stimulated on an isolated leg and the activity in the leg nerve recorded, then it will be found that the
strength of the stimulus diminishes if the stimulus is repeated rapidly, or sustained. This is due to fatigue in the periphery
nerve and sensor, and this ensures that insects respond to changes in the environment, and learn to ignore persistent
stimuli that are of no relevance. For example, body lice prefer rough fabrics, like wool, to smooth fabrics like silk. When
crossing from wool to silk they will keep turning as they try to find their way back onto the wool. However, if they fail to find
the wool (say if it is removed) then this behaviour stops as the lice learn to make do - they habituate. Habituated insects are
still responsive to changes in the stimulus, however. Thus, if the texture of the fabric for our louse on silk changes again,
then we expect it will respond again, and habituate again if necessary.
Learning one's way about - route learning. Some insects, such as ants and cockroaches, are capable of learning the route
of a maze, if the maze exit leads to reward, or if escaping from the maze avoids punishment. A cockroach will navigate a
maze to it home-pot, so long as the home-pot contains recognisable cockroach odours. On subsequent trials it will reach its
home-pot with increasing ease as it gradually learns the route. When compared to rats, ants learn mazes at half the speed:
a rat will master a maze after about 15 runs, an ant after about 30, though the ants still make a few errors. (Not bad when
you consider how tiny the ant brain is!). However, if a change is made to the maze, such as reversing the pattern, then rats
learn the 'new' maze more rapidly than before, recognising the similarity in the patterns and transferring their previous
learning to the new situation (transfer learning). In contrast, the ant starts all-over again, treating the maze as entirely
new. Thus ants have little or no capacity for transfer learning, that is they seem unable to apply what they have learnt to a
novel situation.
One of the most impressive feats of insect learning is locality learning. In addition to learning routes, insects can
recognise the locality in which their nest is situated, or in which food is found. This involves exploratory learning- the insect
will typically fly around a bit after leaving the nest, learning the position of many landmarks very rapidly and then leave. This
is latent learning, meaning that a period of time elapses between learning and reward. The reward occurs when the insect
returns home and locates its nest. This can be demonstrated by experiment. For example, in classic experiments on the
beewolf, Philanthus triangulum a type of digger wasp, which brings back food to its developing young in the nest, the wasp
will learn to recognise a ring of pine cones placed around the nest entrance and that the nest hole is the centre of this ring.
If the pine cones are displaced a few centimetres when the wasp leaves, however, it will return to the centre of the ring of
pine cones (only its nest is no longer there!).Generally, however, these insects are only temporarily confused by changes to
one or two landmarks, falling back on other landmarks further from the nest (and perhaps other cues like smell?).
Associative learning occurs when a stimulus, irrelevant by itself, is made relevant by pairing it with something meaningful,
like a food reward and the animal learns to identify the hitherto unconnected stimulus with food. For example, a cockroach
can be rewarded by being given food when presented with a particular, but unrelated, odour. It will then learn to associate
that odour with food and be attracted to it when foraging. Bees can be trained to associate a particular colour or pattern
9though their perception of shape is limited) with food.
Short-term and long-term memories. When an insect has just completed a task, learning is abolished if a new task is
undertaken immediately after. Learning requires a latent period of rest in-between activities. To some extent the same is
true of humans - learning becomes greatly enhanced if breaks are taken every 20 minutes or so during study. It is during
these rest periods that the brain processes the information (often subconsciously in humans) and the appropriate neural
pathways are reinforced. [Dreams in humans are especially curious - if activities are undertaken straight after waking then
dreams are often very easily forgotten, but if a few minutes are spent reflecting on them they may be remembered more
easily]. This can be explained by the existence of short-term memory (STM) and long-term memory (LTM). The nervous
system requires time to process information in short-term memory, making sense of it and discarding information deemed
irrelevant and transferring more useful information into long-term memory. In insects LTM typically retains information for
several days.
Do Insects Sleep? Hard to say. Insects certainly exhibit circadian cycles. Insects enter a dormant or semi-dormant state
when chilled. This may occur seasonally, during winter, or daily at nightfall. At nighttime insects may assume a particular
posture in which to rest, bees may hold onto vegetation with their tightly clamped jaws, for example. They may also return to
the same resting spot each night. During these quiescent states the body temperature drops and energy is conserved.
Interestingly, learning increases when insects are given rest between tasks or training sessions, as already mentioned, but
also if they are chilled during these rest periods. Could it be that when chilled into torpor at nighttime insects are forming
memories? It has been suggested that two chief functions of sleep are: energy conservation and memory formation. In this
case the insect nighttime torpor is not so very different. Many insects are, of course, nocturnal. Cockroaches are nocturnal
and learn better at night than during the day. It has been suggested that this encourages them to remember useful things
like lessons learned during nighttime foraging.
Temporal learning occurs in insects. As we have seen insects can measure time, by the use of internal body-clocks, and
they can also compensate for the changing position of the Sun in the sky when they use the Sun to navigate. To navigate
successfully by the Sun they must know what time it is, in the sense of what portion of daylight has passed. Observations
have shown that this may require learning, with young insects making mistakes by assuming that the Sun stays fixed in the
sky.
Insight learning is a higher form of learning, similar to transfer learning, in that it takes prior learning and applies that to a
new problem. One example is tool use. Some insects use tools, for example, some digger wasps fill in their nest burrow once
the young is mature and ready to pupate (and so requires protection but no more food). The female may then hold a small
'pebble' in her mandibles and use it to pound-down the earth, and then discard the pebble. However, this behaviour is not
intelligent in the sense that the insect reasoned a solution, rather it is inherited genetically and is instinctive and so not true
insight learning.
The Peripheral Nervous System
Peripheral nerves may be sensory or motory, but in insects are generally mixed. For example, the antennal nerve carries
many sensory fibres conveying inputs from the many antennal sensors to the brain, but it also contains some motor fibres
carrying output signals to the muscles in the base of the antenna. Animal nerve fibres, which are usually the axons of
neurons, are typically wrapped by insulating glial cells. In nonmyelinated axons, the wrapping may be a simple sheath
that loosely invests a group of axons. One function of this sheath is to ensure that the axons are bathed in a suitable salt
solution necessary for them to conduct impulses. In myelinated axons, however, each individual axon is tightly wrapped in its
own insulating sheath of material called myelin, which folds tightly around the axon several times, except at exposed
regions (such as the nodes of Ranvier in mammals, although these may be simple pores in other animals). This more
advanced type of wrapping is insulating and serves to speed-up nerve transmission. Students of vertebrate zoology often
mistakenly believe that only vertebrates possess myelinated axons (in addition to unmyelinated axons which also occur in
vertebrates). Myelinated axons occur in many invertebrates. Indeed, the fastest nerve transmission in the animal kingdom is
seen in the myelinated axons of certain shrimps, which conduct signals at about 200 m/s. Insects have an intermediate form
of insulation, which is like the myelin sheath, except that the myelin is wound loosely, leaving fluid-filled spaces between the
layers. Such a nerve fibre is intermediate between unmyelinated and myelinated and is called a tunicated nerve fibre.
Although Wigglesworth reported seeing the occasional classical myelinated nerve fibre in some insects, they are generally
considered absent from insects. Animals generally invest in the kind of sheathing that is adequate for their purposes. Large
and fast-moving animals, like humans, and small but very fast-moving animals like certain shrimps use myelin where
needed. insects are perhaps too small to require much, if any, use of myelin. The intermediate tunic seems quite sufficient
for their purposes. [A good external link to invertebrate myelin is: http://www5.pbrc.hawaii.edu/~danh/InvertebrateMyelin/ ].
The exact arrangement of ganglia is
grasshopper, left, the first three
abdominal ganglia (A1, A2 and A3) are
fused with the third thoracic ganglion
(T3) to form a composite ganglion. The
final abdominal ganglion is often
composite. It is debatable how many
abdominal segments there are, as the
last few are modified and reduced, but
generally there are 9-10, and the
ganglia of these segments are fused
with that of A8, again forming a
composite ganglion ('A8'). Each
ganglion gives off nerves to the
various structures in its body segment.
However, complications arise as
ganglia may receive inputs from
certain other segments too.
In flies ganglia are highly fused.
Typically the three thoracic ganglia
and the first four abdominal ganglia
are fused together, into a single
ganglion in the thorax. The remaining
5 (or so) abdominal ganglia are also
fused into a single abdominal ganglion.
The connectives of the nerve cord
between the composite thoracic and
composite abdominal ganglia then give
off pairs of nerves to those abdominal
segments lacking a regional ganglion.
Such fusion of ganglia concentrates
processing power where it is needed
and reduces the time wasted by
sending signals up and down the
nerve cord between ganglia that may
need to cooperate.
grasshopper, left, the first three
abdominal ganglia (A1, A2 and A3) are
fused with the third thoracic ganglion
(T3) to form a composite ganglion. The
final abdominal ganglion is often
composite. It is debatable how many
abdominal segments there are, as the
last few are modified and reduced, but
generally there are 9-10, and the
ganglia of these segments are fused
with that of A8, again forming a
composite ganglion ('A8'). Each
ganglion gives off nerves to the
various structures in its body segment.
However, complications arise as
ganglia may receive inputs from
certain other segments too.
In flies ganglia are highly fused.
Typically the three thoracic ganglia
and the first four abdominal ganglia
are fused together, into a single
ganglion in the thorax. The remaining
5 (or so) abdominal ganglia are also
fused into a single abdominal ganglion.
The connectives of the nerve cord
between the composite thoracic and
composite abdominal ganglia then give
off pairs of nerves to those abdominal
segments lacking a regional ganglion.
Such fusion of ganglia concentrates
processing power where it is needed
and reduces the time wasted by
sending signals up and down the
nerve cord between ganglia that may
need to cooperate.
Above: some fibre tracts have been
added: tracts in the optic lobe (notice
the crossovers) and connectives
between pairs of lobes.
added: tracts in the optic lobe (notice
the crossovers) and connectives
between pairs of lobes.
Above: neurosecretory cells with cell
bodies in the pars intercerebralis
added. The axons of these neurons
terminate in the corpora cardiaca
where they releases hormones: part of
the insect neuroendocrine system.
bodies in the pars intercerebralis
added. The axons of these neurons
terminate in the corpora cardiaca
where they releases hormones: part of
the insect neuroendocrine system.
References and Further Reading:
http://casas-lab.irbi.univ-tours.fr/Circadian%20rhythms%20in%20insects.pdf
http://www5.pbrc.hawaii.edu/~danh/InvertebrateMyelin/
Homberg, U. Christensen, T.A. and J.G. Hildebrand, 1989. The structure and function of the deuterocerebrum in insects.
Annu. Rev. Entomol. 34: 477-501.
Howse, P.E. 1975. Brain structure and behaviour in insects. Annu. Rev. Entomol. 20: 359-379.
Fahrbach, S.E. 2006. Structure of the mushroom bodies of the insect brain. Annu. Rev. Entomol. 51:209–32.
http://casas-lab.irbi.univ-tours.fr/Circadian%20rhythms%20in%20insects.pdf
http://www5.pbrc.hawaii.edu/~danh/InvertebrateMyelin/
Homberg, U. Christensen, T.A. and J.G. Hildebrand, 1989. The structure and function of the deuterocerebrum in insects.
Annu. Rev. Entomol. 34: 477-501.
Howse, P.E. 1975. Brain structure and behaviour in insects. Annu. Rev. Entomol. 20: 359-379.
Fahrbach, S.E. 2006. Structure of the mushroom bodies of the insect brain. Annu. Rev. Entomol. 51:209–32.
See also: insect behaviour and control systems; insect society; insect antennae; insect vision.
| More on the insect nervous system: Mechanoreceptors Thermohygroreceptors Chemoreceptors |