The "intelligent" octopus

cartoon of an octopus "The Thinker" courtesy Roland Anderson and Seattle Aquarium, WashingtonThere has been much study of the neural control of behaviour, including learning, in octopuses.  Little comparable work has been done with squids, mainly owing to their mostly pelagic life style and to the difficulty of maintaining them in enclosed spaces.  Only a few learning and related studies have been published on west-coast octopuses.

Octopus "The Thinker" courtesy Roland
Anderson and Seattle Aquarium, Washington

  black dot
  Brain, behaviour, & learning
  The section on the "intelligent" octopus is divided into brain, behaviour, & learning, considered here, and PERSONALITY, SELF-AWARENESS, & THE NEED FOR "ENVIRONMENTAL ENRICHMENT" considered in another part.
Research study 1

drawing of human and octopus eyes to show comparative differences and similaritiescartoon showing comparative brain sizes of comparably sized insect, octopus, and humanMuch has been written on the ability of the European Octopus vulgaris to learn visual and tactile discriminations, to remember these for periods of several weeks, to run mazes, and so on, but there is no demonstration of tool use or of other commonly accepted “higher” indicators of intelligence.  Thus, while octopuses can learn from experience and can retain memories for short duration, there is no ability to respond quickly and successfully to new situations, no sense of knowledge or mental ability, no use of reason to solve problems, and no indication of presence of other generally accepted criteria of intelligence.  They do possess the largest relative brain sizes of any invertebrate, and have large optic ganglia associated with large, image-resolving eyes.

NOTE  an octopus, with appropriate stimulation, learns easily to discriminate between 2 cut-out shapes such as circle and square.  The stimulus to learn comes from being rewarded for responding to the sight of one object and punished with a mild electric shock for the other.  Using the same basic protocol, much finer visual resolutions can be tested, and also touch and chemosensory ones

schematic showing resolving powers of human vs. octopus eyesNOTE based on the density and type of photoreceptors in the retina of an octopus, their resolving power appears to be about one-third that of a human, and they do not see in colour.  Based on recent work on Octopus vulgaris and O. briareus (neither present on the west coast) they have the ability to discriminate images in different degrees of polarised light.  Polarisation vision may be useful in photograph of lens of eye of Humboldt squid Dosidicus gigas showing inverted imagelocating crustacean prey, some species of which are known to reflect light that is strongly polarised, or perhaps, in some unknown way, in communicating with other octopuses.  Shashar & Cronin 1996 J Exp Biol 199: 999; see also Rowell & Wells 1961 J Exp Biol 38: 827, Moody & Parriss 1960 Nature 186: 839,  and Moody 1962 J Exp Biol 39: 21 for information on discrimination of polarised light in European Octopus species.



Lens of eye of Humboldt squid Dosidicus gigas
showing inverted image of the person holding the lens


  black dot
Research study 2

photo/drawing to show the 3 main "lobe systems" in an octopus brainScientists in Britain have collected and dissected the brains of a broad representation of cephalopods, including 39 squids and other non-octopuses, and 23 octopuses.  Two west-coast octopus species (Enteroctopus dofleini and Octopus bimaculatus) and one west-coast squid species (Rossia pacifica) are included.  There are about 30 different lobes to a cephalopod brain, and these can be lumped into a dozen different “lobe systems” that are involved in major sensory and motor functions.  Of the 12, three are of special interest in terms of learning and memory, tactile perception, and vision.  Let’s see where the 3 west-coast cephalopods rank in this world listing of sizes of these 3 systems relative to overall brain sizes.  For comparison, 3 other cephalopods are included along with the west-coast species. These are an unidentified species of open-water squid Loligo sp., the deep-inhabiting giant squid Architeuthis dux, and the European Octopus vulgaris, about which much is known concerning brain and behaviour.

NOTE the data are presented as volume of a certain "lobe system" relative to overall brain volume in % units

1.  VERTICAL-LOBE SYSTEM, comprising superior frontal, vertical, subvertical, and precommissural lobes:  the relative size of this system provides a measure of visual-memory capability.  It would be expected to be larger in shallow bottom-dwelling forms than in deep pelagic forms.


OCTOPUSES (23 listed):
table showing relative size of the vertical-lobe system in several species of octopods

The west-coast giant octopus E. dolfleini does not compare well against other octopods, while the Californian 2-spotted octopus is a veritable “brain". All 3 species are similarly shallow-dwelling and active.

SQUIDS & others (39 species listed):          
table showing relative size of the vertical-lobe system in several species of squids

photograph of stubby squid Rossia pacifica courtesy Roland Anderson and the Seattle Aquarium, WashingtonThe west-coast stubby squid ranks only 17th out of 39 species. Photograph courtesy Roland Anderson and Seattle Aquarium.

2.  INFERIOR FRONTAL-LOBE SYSTEM, comprising superior buccal, posterior buccal, lateral/medial inferior frontal, and subfrontal lobes: the relative size of this system provides a measure of tactile-memory capability.  It is inversely related to the size of the optic lobes (i.e., touch vs. vision).  It would be expected to be relatively larger in bottom-dwelling than in pelagic species (i.e., touch vs. open-water vision).
OCTOPUSES (23 listed):
table showing relative size of the inferior frontal-lobe system in several species of octopods
The giant west-coast octopus ranks a little better than the median, but is well back of the Californian 2-spotted octopus. These 3 species are useful for comparison, as they are ecologically comparable (i.e., shallow-inhabitiing, active).
SQUIDS & others (39 listed): 
table showing relative size of the inferior frontal-lobe system in several species of squids
The west-coast stubby squid ranks comparatively high for potential tactile-memory capability. In comparison with the other squids listed by the authors, which are mostly, or all, open-water species, Rossia lives in close contact with the sea bottom.

3.  OPTIC-LOBE SYSTEM, comprising the large optic lobes and associated parts of brain: provides a general measure of capacity for visual processing.  The size of these lobes is much greater in squids than in octopuses living at comparable depths.  They are also larger in epipelagic (floating above the sea bottom, but generally in deep water) octopuses than in bottom-dwelling ones.

NOTE  the optic lobes are generally bigger in volume than the brain, in this case by an overall average for the 39 species of 1.2 times

OCTOPUSES (23 listed):
table showing relative size of the inferior optic-lobe system in several species of octopods
Neither west-coast species ranks particularly high for this attribute.  In the case of Enteroctopus dofleini, a species that inhabits murky water and is active at night, it may indicate less reliance on vision and perhaps more reliance on tasting and feeling its way around its habitat.
SQUIDS & others (39 listed):
table showing relative size of the optic-lobe system in several species of squids  
The shallow, bottom-dwelling stubby squid ranks surprisingly low in potential capacity for visual processing, as does the giant squid Architeuthis, which lives more-or-less continuously in perpetual blackness.

Although there is nothing “hard and fast” about these values, they suggest that in comparison with other world species of cephalopods, Enteroctopus, albeit large and robust, could be a bit of a mental dummy, with low-ranking visual-memory and visual-processing capability, and with only better than average tactile-memory capabilities.  In comparison, the smaller Californian 2-spotted octopus O. bimaculatus is potentially much sharper mentally, with possibly greater visual and tactile memory systems, and much better visual-processing capability.  Maddock & Young 1987 J Zool, Lond 212: 739.

Research study 3

In the laboratory, octopuses Octopus rubescens will readily attack crabs Hemigrapsus oregonensis impaled upon a rod and displayed in front of them.  If black or white rectangular objects1 are presented with the crab, and the octopus is rewarded2 for attacking in one circumstance, say, when the black object is presented, and punished for attacking in the other circumstance, it can be trained to discriminate between the two.  However, the training does not come easy for O. rubescens, at least not in comparison with O. vulgaris from Europe on which many learning experiments have been done.  For example, a simple black/white visual discrimination of the type3 described that O. vulgaris can learn with 100% fidelity in about a week and remember mistake-free for a month, will take O. rubescens about 40wk, and then the proficiency attained is only 7 out of 8 correct responses over a 3-d consecutive period.  Warren et al. 1974 Anim Behav 22: 211.

NOTE1  these are solid plastic, 15 x 5 x 5cm in size, mounted on a rod for a handle, and bearing a nail-like fixture projecting downwards on which a crab can  be impaled. The octopus is rewarded when it makes a correct choice

NOTE2  reward is the crab, while punishment is a mild electric shock (8V DC) applied with prongs

NOTE3  8 trials per day comprising 4 rewards (when black is shown) and 4 punishments (when white is shown)

Octopus rubescens
stranded in a tidepool at
Botanical Beach, British Columbia 0.8X

Research study 4

graph showing times for octopuses Octopus dofleini to open jars containing prey over successive days photograph of octopus Enteroctopus dofleini with bottle used in training testsOctopuses have long been thought to be incapable of using their arms and suckers for manipulative tasks, such as removing a cork stopper from a glass jar to capture a crab inside.  That they occasionally do so was credited to happenstance.  At least part of the reason for this is that octopuses seem to lack the proprioceptive ability to assess the positions of their arms in space.  So, removing a screw-top lid from a jar to capture a crustacean inside would not be possible…or would it?  Perhaps the appropriate stimuli have never been presented.  In a new look at an old problem, researchers at the Seattle Aquarium present octopuses Enteroctopus dolfleini with clear plastic screw-top jars, each with a single crayfish inside, in 3 experimental configurations: 1) closed jar with visual cues only, 2) minimal chemical cues provided by six 1cm holes being drilled in the jar, and 3) maximal chemical cues provided by herring mucus being smeared on the outside of a drilled jar.   A 60-min time limit is arbitrarily set, after which a “no response” is scored. 

Results show that for configurations 1 & 2, only 3 jars are opened out of a total of 20 trials (10 trials for each).  However, with the added stimulus of the herring slime, an average of 9.5/10 jars are opened.  Average time taken for 1 & 2 is about 18min, as compared with 41min for configuration 3.  So, we know that an octopus will fiddle with its jar more owing to the added stimulus, but where is the learning component?  This comes with repetitive testing of the same individuals, with results for configuration 3 shown in the accompanying graph. Note that opening times drop significantly from a mean of 41min on Day 1 to 15min on Day3, indicative of learning.  If these trained individuals are now re-presented with type-1 jars (no holes, no herring mucus), their jar-opening performance is significantly faster (data not provided by the authors). This is a marvelous contribution to the study of learning and behaviour in octopuses, and is sure to generate a host of follow-up research questions. Anderson & Mather 2010 Ferrantia 59: 8. Photograph courtesy the authors and the Seattle Aquarium.

NOTE  owing to a limit on the number of octopuses available for study, most or all of 11 individuals are used more than once (in varying order of testing).  Ideally, of course, one octopus should be used for one trial, then “discarded”, with the exception of the learning part of the experiments, where an individual is required to be tested repetitively

NOTE  each jar with crayfish is adjusted such that a single revolution is required to remove the lid