Cross-posted on the Animal Ethics website here. This review was written as part of a volunteer internship I've been doing with AE over the last few months.

Introduction

Most of the animals on our planet are invertebrates. The invertebrate label applies to a hugely diverse range of animals, comprising 99% percent of all species (Ray, 2018) and 99.9998% of all animals (Bar-On et al., 2018). This includes the 139 µm long parasitic wasp, Dicopomorpha echmepterygis, as well as the 15m squid, Mesonychoteuthis. Such diversity makes evaluating sentience in invertebrates challenging, and research is crucial due to the enormous number of individuals who may be sentient and experiencing harms.

Human activities impact invertebrates in numerous and significant ways – mussels are harvested for food, insects are killed by our pesticides, lobsters are boiled alive, and octopodes are kept in aquariums, to name just a few common human-invertebrate interactions. In the wild, invertebrates are subject to situations including disease, death by starvation and predation. It is clear that many of the experiences which invertebrates are subject to would elicit suffering in humans or other sentient vertebrates. Knowledge of which invertebrates have subjective experience of the world would have important implications for human actions which affect invertebrates under our direct control. It would also have implications for actions which affect invertebrates in the wild. Despite the importance of such research, at present relatively little research has been done on the question of invertebrate sentience.

This literature review focuses on what the neuroscientific evidence suggests regarding invertebrate sentience (the words sentience and consciousness are used interchangeably in this review).^[1] Whilst behavioural evidence can also be useful for making inferences about the presence (or not) of sentience in animals, it can be difficult to know whether observed behaviours signal consciousness or only an unconscious programmed response (similar to computer algorithms that exhibit complex ‘behaviours’ such as playing a game of chess, without consciousness necessarily being involved). Insights from neuroscience are useful because they enable us to understand the physical structures that underlie consciousness in the brain.

There are many theories of consciousness and subscription to any one of these theories can change the probabilities one assigns to the presence of consciousness in different nonhuman animals. This review will be, as far as possible, theory-neutral in this respect. Thus, the evidence presented for and against invertebrate sentience should be compelling to a range of theoretical perspectives. This review does not examine every neuroscientific feature which is relevant to sentience,^[2] but instead focuses on three features which seem particularly important: number of neurons in the brain, presence of a specific brain structure (or its functional equivalent), and degree of centralisation. Therefore, this review should be viewed as only a snapshot of the neuroscientific research related to invertebrate sentience.

Neuroscientific features relevant to sentience

1. Number of neurons in the brain

Neurons are the basic building blocks of all biological brains. Neurons in the brain are connected to each other, forming networks where information can be transmitted, processed, and integrated. This handling of information by the network enables organisms to respond to stimuli in their environment with appropriate behaviours, a function essential for survival. The number of neurons present in the brains of different invertebrates varies by many orders of magnitude. The hermaphrodite of the small nematode worm C. elegans contains just 302 neurons (Hobert, 2018), with only a subset of these comprising the brain. The brain of the sea slug Aplysia californica contains around 3000 neurons (Cash & Carew, 1989) and the fruit fly brain has about 100,000 neurons (Chiang et al., 2011), whilst the common octopus Octopus vulgaris has approximately 45 million cells in the central brain complex (Hochner, Shomrat, & Fiorito, 2006). For reference, the human brain contains approximately 86 billion neurons (Herculano-Houzel, 2012).

The literature shows that behaviour and cognition beyond basic reflexes can be mediated by networks of even a small number of neurons. For example, nematodes are capable of performing mating behaviour requiring precise manoeuvres (Barr, 2006), and display associative learning and long-term memory capabilities (Ardiel & Rankin, 2010).^[3] Computational models of neural networks have supported the idea that surprisingly advanced behavioural and cognitive processes can be performed with only minimal neural circuitry. Chittka & Niven (2009), for example, point out that a simple visual categorisation task can be performed with a network of 12 neurons, or that basic numerical abilities can be simulated using a few hundred units. Chittka & Niven note that the hypothesis that brain size (which correlates with neuron number) is a good predictor of behavioural repertoire and cognitive capacity is controversial. A striking comparison is made by the authors between the moose and the honeybee, which demonstrates the unreliability of brain size in predicting these abilities: moose exhibit 22 distinct behaviours compared to bees’ 59, despite the obviously larger moose brain (Chittka & Niven, 2009). However, it should be noted that the ethograms used to catalogue these behaviours are not necessarily well-standardised and this difference could be due to differences in methodology.^[4]

Chittka & Niven (2009) offer a reason for why number of neurons is unlikely to have much of an impact on the kinds of processes occurring in a brain: larger brains are likely to be in larger bodies. These larger bodies require greater replication of neuronal circuits in order to add precision to sensory processes, detail to sensory perception, and increase storage capacity: all things which don’t necessarily produce qualitative shifts in behaviour or cognition. The authors conclude that qualitative changes in behavioural performance are therefore much more likely to be mediated by the recruitment of new neurons into novel pathways and brain regions (important for greater serial and parallel processing of information and more links between processing pathways), rather than a greater number of neurons.

We may plausibly believe that cognitive and behavioural ability are useful indicators of sentience in nonhuman animals. The evidence presented above suggests that the quantity of neurons in a given brain is not necessarily a good indicator of cognitive and behavioural ability. So, by extension, we may believe it is not a good correlate of sentience either. Indeed, Klein & Barron (2016) make this argument regarding consciousness. They claim that functional organisation is what is important, and neuron number only matters insofar as it affects functional organisation. For them, this means that brains as small as those of an insect have enough neurons to support consciousness.

The point made by Klein & Barron (2016) about the effect of neuron number on functional organisation is particularly important to consider when thinking about organisms that possess very small numbers of neurons. This is because, although Klein & Barron assume that one million neurons in (e.g.) a honeybee can instantiate enough of the functional organisation required for consciousness, such assumptions become increasingly more difficult to justify as the number of neurons being considered decreases. Because we don’t know what kind of neural organisation is required for consciousness, we can currently only guess at the minimum number of neurons needed to instantiate it.^[5] What kind of number seems plausible largely depends on your preferred neuroscientific theory of consciousness.^[6] Among some current popular neuroscientific theories of consciousness, Information Integration Theory (Tononi, Boly, Massimini, & Koch, 2016) would suggest that a very low number of neurons is required for consciousness, whilst Global Workspace Theory would suggest a larger number of neurons are required (Baars, 2005).^[7]

The recognised significance of neural organisation, alongside the uncertainty over what kind(s) of neural organisation support consciousness, demonstrates the importance of investigating the question further. The following two sections examine a couple of the ways in which neural organisation could be important to sentience.

2. Presence of a specific brain structure (or its functional equivalent)

This section examines two separate but related debates. The first concerns the extent to which possession of a cortex or cortex-like structure is required for sentience. The second concerns whether consciousness can be supported by a midbrain or midbrain-like structure.

2.1. Cortex and cortex-like structures

The cortex is the outer layer of neural tissue in the cerebrum of humans and other mammals.^[8] It is an important site of neural integration in the brains of those animals and has been implicated in a variety of higher-level functions, including memory, perception, and attention. Historically, the cortex has been viewed by many as integral to the generation of consciousness, and, according to Merker (2007) this view continues to hold sway.^[9]

The traditional claim that the cortex is necessary for consciousness, at least in humans, is founded partly on the belief that absence or destruction of cortical regions abolishes consciousness, such as in hydranencephalic patients who are born without a cortex (Hill, 2016), or in people which lose cortical function after brain damage (Puccetti, 1988). The claim that we are unaware of all activity occurring in the brainstem and spinal cord (Rose, 2002) is another argument for the necessity of a cortex for consciousness. Key (2016) argues that certain types of organisation – which are found in the cortex (or its functional equivalents) – are necessary for subjective experience, including: the formation of sensory maps, lamination allowing for complex wiring, multiple layers, and strong local and long-range connections. Such cortex-focused views therefore lead to conclusions such as those of Edelman et al. (2005), who argue that possession of “neural structures that are functional equivalents of cortex and thalamus [a small structure often considered important due to the strong reciprocal connections it has with the cortex]” is necessary for consciousness in non-mammalian species.

Requirements such those listed by Key (2016) are a major challenge to the idea of consciousness in invertebrates (and indeed, all non-mammals), since they set the bar high for non-corticate animals. Despite this, the case for possession of brain regions homologous to cortex has been made quite successfully for birds (Harris, 2015). It has also been suggested that octopodes possess brain regions which work in ways similar to cortex: Edelman & Seth (2009) present the vertical, superior and optic lobes of the octopus as possible sites for homologous “low-amplitude fast” neural activity, which they consider a key feature of mammalian consciousness. This suggests that activity functionally equivalent to that occurring in the mammalian thalamocortical complex is possible in at least one invertebrate.^[10] However, even if a case can be made for some animals outside of the mammalian class, a view which requires cortex or a functional equivalent for sentience still rules out the possibility of consciousness in most invertebrates. This is because the majority of invertebrates lack any brain structures remotely similar to a cortex (Barron & Klein, 2016).

The discussion about whether a cortex is required for consciousness now leans in favour of the argument that it is not. The most significant development in recent years has been the release of “The Cambridge Declaration of Consciousness”, a statement signed by a group of prominent neuroscientists which states that “the absence of a neocortex does not appear to preclude an organism from experiencing affective states” (Low, 2012). The signatories therefore conclude that not only mammals, but also birds and other animals such as octopodes possess the necessary neurological substrates to generate consciousness. In his report on consciousness, Muehlhauser (2018) also rejects what he terms the “cortex-required view”, concluding that there is not currently even a moderately strong case for the cortex being a necessary condition for phenomenal consciousness.

2.2. Midbrain and midbrain-like structures

Some researchers have argued against the view that a cortex or its functional equivalent is necessary for consciousness even in humans and other mammals. This debate is important because the argument entails that the midbrain would be sufficient to support consciousness. Although sentience in most invertebrates is doubtful on a cortico-centric view of consciousness, some invertebrates may be considered sentient under this more inclusive midbrain-centric view.^[11]

One of the main opponents to the cortico-centric view, Bjorn Merker, has argued that the key mechanisms of consciousness are implemented in the midbrain and basal diencephalon of mammals, rather than the cortex (Merker, 2007). One piece of evidence in favour of this argument is the observation that hydranencephalic children are capable of displaying behaviours indicative of consciousness, such as emotions and absence epilepsy (a condition which is typically an affliction of the conscious state).^[12] Another line of evidence in favour of Merker’s argument comes from studies in anaesthesia. Structures in the brainstem and diencephalon were found by Mashour & Alkire (2013) to be sufficient to support primitive consciousness in humans, with “only limited neocortical involvement.”^[13] Merker (2007) concludes that the brainstem (a region including the midbrain) is sufficient for mediating consciousness.^[14] Other researchers have also supported the view that the cortex is not necessary for conscious experience: for example, Panksepp (2011) concludes that subcortical structures in mammals are sufficient for the generation of emotions.

The claim that the midbrain is capable of supporting conscious experience in mammals is contentious and has attracted a fair amount of criticism. In response to claims that hydranencephalic patients are conscious purely by virtue of their subcortical activity,^[15] Watkins & Rees (2007) point out that a majority of these patients have some cortex remaining and that the heterogeneity of the condition means that, without more careful analysis, it is difficult to derive the conclusion that consciousness is supported without the presence of any cortex.^[16] Allen-Hermanson (2016) observes that although many recent papers have acknowledged the importance of subcortical structures, they generally also consider the cortex to play a key role in consciousness and rarely unequivocally conclude that subcortical regions are sufficient. For example, Damasio & Carvalho (2013) write that subcortical structures are a “candidate” for the neural substrate of feelings, but in the same paragraph note that cortical regions are also candidates.

The debate over whether midbrain or midbrain-like structures are capable of supporting conscious experience is currently inconclusive. The outcome of this debate is significant to the question of sentience in many invertebrates. This is because, if the midbrain-sufficient view is correct, it enables the argument that some invertebrates which don’t possess a cortex-like structure are conscious because they possess midbrain-like structures. For example, arguments for consciousness in phyla such as Arthropoda can be developed on the basis of similarities between their brain structures and those of the mammalian midbrain. This line of thinking has been deployed by Barron & Klein (2016) to argue for consciousness in insects.

2.3. Other structures capable of supporting consciousness

Of course, consciousness need not necessarily be instantiated in invertebrates by any structure which is recognisably similar to a structure in the mammalian brain. It is plausible that consciousness can be supported by many different varieties of neural organisation.^[17] Progress in understanding the types of organisation which give rise to sentience will be important for better understanding the likelihood of sentience in invertebrates (and all other animals). Progress will be particularly important for understanding the likelihood of sentience in simpler invertebrates such as nematodes and bivalves. This is because the nervous system anatomy in these animals, due to its relative simplicity and small size, is radically different from anything found in mammals (or any other vertebrate). As a result, comparisons with the mammalian cortex or midbrain are more difficult and an understanding of the role of particular processes becomes more useful.

3. Degree of centralisation

3.1. Links between centralisation and consciousness

Integration of information into a coherent model of the world is an important aspect of many views about the function of consciousness in animals (Baars, 2005; Feinberg & Mallatt, 2013; Merker, 2005; Morsella, 2005). The link between consciousness and integration of sensory information into a single, unified experience is a reason to think that the degree of centralisation in a nervous system is important to the generation of conscious experience (Animal Ethics, 2014). This is because it seems likely that a higher degree of centralisation in a nervous system allows for closer integration of disparate sensory information into a unified whole. If there were not a sufficient degree of centralisation, we could imagine that the streams of information from different sensory modalities and/or parts of the body would remain somewhat separate. So, if we believe that unified experience is a key aspect of consciousness, then we might think that there is no conscious experience in a non-centralised case.

The different circuits involved in conscious and unconscious processing in the human nervous system represents a second reason to think degree of centralisation is important to consciousness. Whilst we know that some of the processes which occur in the human brain are conscious, information processing which takes place in local circuits in the peripheral nervous system (PNS) is unconscious.^[18] Local circuits in the PNS operate outside of the central nervous system and are associated with reflex actions such as the patellar reflex (the knee-jerk response), which is triggered by the interaction of only two neurons. Since the highly centralised and integrative circuits in the brain give rise to consciousness in humans, whereas the more diffuse circuits in the PNS do not,^[19] we have another reason to believe that the degree of centralisation is an important factor in whether a particular nervous system organisation supports consciousness.

We do not know what degree of centralisation is required for consciousness. However, thinking about the functionality of nervous systems with different degrees of centralisation is useful when considering the question. If a nervous system is acting as a simple stimulus-response mechanism that does not generate an egocentric model of the world, it is likely operating without consciousness. However, if a nervous system is capable of combining different types of sensory information into a map of the world and then selecting an appropriate behavioural response based on an evaluation of its current motivations – then the system might well be conscious. Relatively de-centralised systems such as those found in bivalves tend to show a little evidence of the former mode of function.^[20] In contrast, more centralised systems such as those in insects show some evidence of the latter mode of function.^[21] However, factors other than degree of centralisation, such as the number of neurons, also affect how capable a nervous system is of integrating spatial information into a unified sensory map. For example, despite being highly centralised, the small size of the nematode nervous system might mean it is not capable of generating an egocentric sensory map of the world (Barron & Klein, 2016).

3.2. Centralisation in different nervous systems

Broadly, vertebrate nervous systems are characterised by greater centralisation of information processing, and invertebrates by less centralisation (although with significant variance between phyla). This is primarily because all vertebrates possess brains that are distinct from the rest of their nervous systems and which clearly perform the role of information integration and action selection whereas, among invertebrates, the presence of a distinct brain region can often be less apparent. Additionally, invertebrates mostly have unipolar neurons whereas vertebrates mostly have multipolar neurons, which contain branching axons and multiple dendrites connected to the cell body, potentially allowing for more efficient integration of information between neurons (Smarandache-Wellmann, 2016). Some invertebrates such as insects do contain concentrations of neurons which clearly perform an executive function (Gronenberg & López-Riquelme, 2004), which we would label as a brain.^[22] However, among some invertebrates, the nervous system more closely resembles a collection of “ganglion” units distributed throughout the body, such as in bivalves (Thorp, 1991). In invertebrates with some of the simplest nervous systems, such as Cnidaria (a phylum which includes jellyfish), there is debate as to whether there is centralised processing at all, since the neuroanatomy is mainly characterised by nerve net structures (Satterlie, 2011).

Although there is no agreed-upon method for measuring how centralised the information processing in a particular nervous system is, it is possible to guess at some criteria which plausibly affect the degree of centralisation in a nervous system. Identifying clustering of neurons is the most obvious place to begin and is the criterion according to which invertebrate nervous systems have been roughly sorted into three types in the previous paragraph. In addition, interneuronal distance, axonal conduction velocity, and synaptic transmission velocity would affect the speed at which information is processed, contributing to a greater or lesser amount of information integration per time unit. Furthermore, the extent of cortical folding and the average number of connections per neuron affect the total connectivity in a cluster of neurons. Again, this would impact the quantity of information which can be integrated by a system in a given time period. A review paper by Dicke & Roth (2016) lists many of these variables as being important for information processing capacity, which they link to general intelligence. This strengthens the case that these factors may also contribute to the generation of conscious experience.

It should be kept in mind that whilst degree of centralisation (as defined by some quantitative measure) is likely relevant to the capacity of a nervous system to support consciousness, there are also qualitative variables which should always be considered in tandem. For example, it might matter not just how much information a network is processing, but also what kind of information (sensory data? emotions? higher-order thoughts? etc.) is being processed and in what way that information is being processed (for example, recurrent processing or synchronised firing of groups of neurons might be important). A highly centralised system might process information in a way that does not support consciousness, whereas a relatively more distributed system might be organised in a manner better suited to the generation of conscious experience. Therefore, as emphasised in the earlier sections, qualitative as well as quantitative criteria may need to be met in order for the processes in a nervous system to generate consciousness.

4. Conclusion

A summary of the findings from each of the previous sections is presented below, followed by a discussion of the overall state of the literature.

4.1. Number of neurons in the brain

Although the literature concerning the relationship between number of neurons in the brain and the presence of sentience is sparse, there are a couple of points regarding this metric which we can usefully extract. Firstly, the general consensus is that the functional organisation of the neurons in the brain is most important, and that the number of neurons is important primarily in that it affects how complex this organisation can be. Secondly, it is not clear at what point the number of neurons present in a nervous system becomes too low to support a sufficiently complex functional organisation for consciousness.

4.2. Presence of a specific brain structure (or its functional equivalent)

It is now recognised that non-mammals can possess brain structures which function similarly to a mammalian cortex, meaning that the case for consciousness in many other animals, notably birds and octopodes, has strengthened. The argument that the midbrain is sufficient to support conscious experience is not currently widely accepted, but the belief that the cortex is essential to the generation of consciousness has increasingly been challenged in recent years. The focus of the current literature (on finding similarities between the brain structures of nonhuman animals and the brain structures which we believe to support consciousness in humans) is partly a result of our current lack of specific understanding about which types of neural organisation are important for sentience. We must remain open-minded about the possibility that many other (possibly radically different) types of neural organisation could also support consciousness. Research in this area is particularly important for understanding whether consciousness can be supported by the nervous systems of simpler invertebrates.

4.3. Degree of centralisation

There are reasons grounded in both theory and real-world observation for why we might believe that the degree of centralisation in a nervous system is important to whether it can support consciousness or not. There is currently no understanding of how centralised a nervous system must be in order for generation of conscious experience to be possible. Furthermore, little work has been done on how best to measure the degree of centralisation in different nervous systems. Metrics such as neuronal processing speeds and network connectivity may be useful in providing a rough estimate of how much information integration occurs (per unit time) in a given area of a nervous system. However, it is certain that many other factors are also relevant.

It is important that discussion about centralisation of nervous systems happens whilst other relevant factors are kept in mind, since even a highly centralised network of neurons could still lack the size or correct functional organisation to support consciousness. Work on centralisation is particularly important to invertebrate sentience research since the degree of centralisation in the nervous systems of invertebrates ranges much more widely than in vertebrates.

4.4. State of the neuroscientific literature and scope for future research

Academic research on the neuroscientific questions relevant to invertebrate sentience is generally sparse. Many of the papers which have proved useful when researching this literature review do not directly address the topic of invertebrate sentience. Rather, a lot of the papers which are currently most valuable to a researcher of invertebrate sentience are focused on associated topics such as cognitive ability in animals or the underpinnings of human consciousness. In recent years, there has been publication of some direct work on the neuroscience relevant to invertebrate sentience.^[23] However, much further work in neuroscience (as well as other relevant fields) is required in order to make advances on this important topic. The kind of work which might be useful can be broken down into three different categories:

1. What criteria are relevant to invertebrate sentience, and to what degree are they significant?
2. How do different invertebrates stack up relative to these criteria?
3. What practical consequences should our knowledge of invertebrate sentience have?

Whilst this literature review has focused on summarising current neuroscientific knowledge about (1), the state of the neuroscientific literature relevant to (2) is much the same as just noted (i.e. most studies on the nervous systems of specific invertebrates are not currently done with the primary aim of furthering our understanding of sentience, although they can still provide useful information for that purpose). Whilst (3) could partly be considered a philosophical question, neuroscience can play an important role in answering it. This is because the study of invertebrate nervous systems could give us clues about not just whether different invertebrates have experience of the world, but also what kind of experiences they have. For example, we might discover that an invertebrate is conscious of the world but does not have the neural machinery to support certain feelings. Alternatively, we might discover that, for a unit of objective time, some invertebrate’s experience of suffering lasts subjectively longer than a human’s experience would,^[24] in which case we would likely adjust upwards the moral weight which we assign to that invertebrate.

Regarding the three particular criteria discussed in this literature review, further research would be instrumental in answering some of the many questions raised. Improving our understanding of the relative importance of each of the criteria will help us to gain a fuller picture of which invertebrates might be sentient. In doing so, it seems sensible to research the relevant questions from the top-down, by developing theories about the function and origins of consciousness, at the same time as the bottom-up, by seeking better understanding of which types of neural organisation are associated with consciousness. At present, our lack of understanding about which invertebrates are sentient presents a major barrier to actions that might help improve their welfare. Progress in the relevant neuroscience, along with other fields connected to invertebrate sentience, is therefore essential if we wish to care properly for all sentient beings, large or small.

This review was written by Jamie Gittins as part of a volunteer internship with Animal Ethics. I appreciate the support of my supervisor, Oscar Horta. I am also grateful to Jason Schukraft, Leah McKelvie, Max Carpendale and Cyndi Rook for the useful feedback they have provided.

Bibliography

Allen-Hermanson, S. (2016) “Is cortex necessary?”, Animal Sentience, 1 (9), a6.

Animal Ethics (2014) “Animal sentience”, https://www.animal-ethics.org/sentience-section/animal-sentience/.

Ardiel, E. L. & Rankin, C. H. (2010) “An elegant mind: Learning and memory in Caenorhabditis elegans”, Learning & Memory, 17, pp. 191-201.

Baars, B. J. (2005) “Global workspace theory of consciousness: Toward a cognitive neuroscience of human experience”, Progress in Brain Research, 150, pp. 45-53.

Barr, M. M. & Garcia, L. R. (2006) “Male mating behavior”, in The C. Elegans Research Community, Wormbook (ed.) Wormbook, Pasadena: California Institute of Technology.

Barron, A. B. & Klein, C. (2016) “What insects can tell us about the origins of consciousness”, Proceedings of the National Academy of Sciences, 113, pp. 4900-4908.

Bar-On YM, Phillips R, Milo R. The biomass distribution on Earth. Proc Natl Acad Sci. 2018 Jun 19;115(25):6506–11.

Bowmaker, J. K. (1998) “Evolution of colour vision in vertebrates”, Eye, 12, pp. 541-547.

Cash, D. & Carew, T. J. (1989) “A quantitative analysis of the development of the central nervous system in juvenile Aplysia californica”, Journal of Neurobiology, 20, pp. 25-47.

Changizi, M. A. (2003) The brain from 25,000 feet: High level explorations of brain complexity, perception, induction and vagueness, Dordrecht: Springer. Chiang, A.-S.; Lin, C.-Y.; Chuang, C.-C.; Chang, H.-M.; Hsieh, C.-H.; Yeh, C.-W.; Shih, C.-T.; Wu, J.-J.; Wang, G.-T.; Chen, Y.-C.; Wu, C.-C.; Chen, G.-Y.; Ching, Y.-T.; Lee, P.-C.; Lin, C.-Y.; Lin, H.-H.; Wu, C.-C.; Hsu, H.-W.; Huang, Y.-A.; Chen, J.-Y.; Chiang, H.-J.; Lu, C.-F.; Ni, R.-F.; Yeh, C.-Y. & Hwang, J.-K. (2011) “Three-dimensional reconstruction of brain-wide wiring networks in drosophila at single-cell resolution”, Current Biology, 21, pp. 1-11.

Chittka, L. & Niven, J. (2009) “Are bigger brains better?”, Current Biology, 19, pp. R995-R1008.

Damasio, A. & Carvalho, G. B. (2013) “The nature of feelings: Evolutionary and neurobiological origins”, Nature Reviews Neuroscience, 14, pp. 143-152. Dicke, U. & Roth, G. (2016) “Neuronal factors determining high intelligence”, Philosophical Transactions of the Royal Society B: Biological Sciences, 371.

Dugas-Ford, J.; Rowell, J. J. & Ragsdale, C. W. (2012) “Cell-type homologies and the origins of the neocortex”, Proceedings of the National Academy of Sciences of the United States of America, 109, pp. 16974-16979.

Edelman, D. B.; Baars, B. J. & Seth, A. K. (2005) “Identifying hallmarks of consciousness in non-mammalian species”, Conscious and Cognition, 14, pp. 169-187

Edelman, D. B. & Seth, A. K. (2009) “Animal consciousness: A synthetic approach”, Trends in Neurosciences, 32, pp. 476-484.

Feinberg, T. E. & Mallatt, J. (2013) “The evolutionary and genetic origins of consciousness in the Cambrian Period over 500 million years ago”, Frontiers in Psychology, 04 October.

Fiorito, G.; Affuso, A.; Basil, J.; Cole, A.; de Girolamo, P.; D’Angelo, L.; Dickel, L.; Gestal, C.; Grasso, F.; Kuba, M.; Mark, F.; Melillo, D.; Osorio, D.; Perkins, K.; Ponte, G.; Shashar, N.; Smith, D.; Smith, J. & Andrews, P. L. (2015) “Guidelines for the care and welfare of cephalopods in research –A consensus based on an initiative by CephRes, FELASA and the Boyd Group”, Laboratory Animals, 49, suppl. 2, p. 1-90.

Gronenberg, W. & López-Riquelme, G. O. (2004). “Multisensory convergence in the mushroom bodies of ants and bees”, Acta Biologica Hungarica, 55, pp. 31-37. Harris, K. D. (2015) “Cortical computation in mammals and birds”, Proceedings of the National Academy of Sciences, 112, pp. 3184-3185.

Herculano-Houzel, S. (2012) “The remarkable, yet not extraordinary, human brain as a scaled-up primate brain and its associated cost”, Proceedings of the National Academy of Sciences, 109, suppl. 1, pp. 10661-10668.

Hill, C. S. (2016) “Insects: Still looking like zombies”, Animal Sentience, 1 (9), a20.

Hobert, O. (2010) “Neurogenesis in the nematode Caenorhabditis elegans”, in The C. Elegans Research Community, Wormbook (ed.) Wormbook, Pasadena: California Institute of Technology.

Hochner, B.; Shomrat, T. & Fiorito, G. (2006) “The octopus: A model for a comparative analysis of the evolution of learning and memory mechanisms”, The Biological Bulletin, 210, pp. 308-317.

Husband, S. A. (2017) “Of cortex and consciousness: ‘Phenomenal,’ ‘access,’ or otherwise”, Animal Sentience, 2 (13), a7.

Key, B. (2016) “Why fish do not feel pain”, Animal Sentience, 1 (3), a1.

Klein, C. & Barron, A. B. (2016a) “Insect consciousness: Commitments, conflicts and consequences”, Animal Sentience, 1 (9), a21.

Klein, C. & Barron, A. B. (2016b) “Insects have the capacity for subjective experience”, Animal Sentience, 1 (9), a1.

Low, P. (2012) “The Cambridge declaration on consciousness”, http://fcmconference.org/img/CambridgeDeclarationOnConsciousness.pdf.

Mallatt, J. & Feinberg, T. E. (2016) “Insect consciousness: Fine-tuning the hypothesis”, Animal Sentience, 1 (9), a10.

Mashour, G. A., & Alkire, M. T. (2013) “Evolution of consciousness: Phylogeny, ontogeny, and emergence from general anesthesia”, Proceedings of the National Academy of Sciences, 110, suppl. 2, pp. 10357-10364.

Merker, B. (2005) “The liabilities of mobility: A selection pressure for the transition to consciousness in animal evolution”, Consciousness and Cognition, 14, pp. 89-114.

Merker, B. (2007) “Consciousness without a cerebral cortex: A challenge for neuroscience and medicine”, Behavioral and Brain Sciences, 30, pp. 63-81.

Morsella, E. (2005) “The function of phenomenal states: Supramodular interaction theory”, Psychological Review, 112, pp. 1000-1021.

Muehlhauser, L. (2017) “2017 report on consciousness and moral patienthood”, Open Philanthropy Project, June.

Panksepp, J. (2011) “Cross-species affective neuroscience decoding of the primal affective experiences of humans and related animals”, PLoS ONE, 6 (9).

Puccetti, R. (1988) “Does anyone survive neocortical death?”, in Zaner, R. M. (ed.) Death: Beyond whole-brain criteria, Dordrecht: Springer, pp. 75-90.

Ray, G. (2018) “Invertebrate sentience: Urgent but understudied”, Wild-Animal Suffering Research, January 19.

Rose, J. D. (2002) “The neurobehavioral nature of fishes and the question of awareness and pain”, Reviews in Fisheries Science, 10, pp. 1-38.

Satterlie, R. A. (2011) “Do jellyfish have central nervous systems?”, Journal of Experimental Biology, 214, pp. 1215-1223.

Shigeno, S.; Andrews, P. L. R.; Ponte, G. & Fiorito, G. (2018) “Cephalopod brains: An overview of current knowledge to facilitate comparison with vertebrates”, Frontiers in Physiology, 9.

Sumbre, G.; Gutfreund, Y.; Fiorito, G.; Flash, T. & Hochner, B. (2001) “Control of octopus arm extension by a peripheral motor program”, Science, 293, pp. 1845-1848.

Thorp, J. H. (1991) Ecology and classification of North American freshwater invertebrates, San Diego: Academic Press.

Tomasik, B. (2016) “Do smaller animals have faster subjective experiences?”, Essays on Reducing Suffering, Jul 21.

Tononi, G.; Boly, M.; Massimini, M. & Koch, C. (2016) “Integrated information theory: From consciousness to its physical substrate”, Nature Reviews Neuroscience, 17, pp. 450-461.

Tye, M. (2017) Tense bees and shell-shocked crabs: Are animals conscious?, Oxford: Oxford University Press.

Waldhorn, D. R. (2019) “Invertebrate sentience table”, Rethink Priorities, June 14.

Watkins, S. & Rees, G. (2007) “The human superior colliculus: Neither necessary, nor sufficient for consciousness?”, Behavioral and Brain Sciences, 30, p. 108.

Footnotes

---

1. A simple definition of consciousness will suffice for the purposes of this review: it is subjective experience of something it is like to be. ↩︎

2. There are many more specific features of nervous systems which are also plausibly relevant to sentience (e.g. re-entrant processing, neural assemblies) that I have not examined here. Nociception is not discussed because its scope overlaps significantly with other fields outside of neuroscience. ↩︎

3. Memory is “long-term” in the context of a nematode lifespan, i.e. days. ↩︎

4. See Changizi (2003, p. 43) for a table detailing the number of distinct behaviours for 51 species. It contains some surprises, such as a Caiman reptile having 1/3 more distinct behaviours than a human child, suggesting ethograms might have limited usefulness for determining the cognitive capacities of animals. ↩︎

5. Luke Muehlhauser (2018) writes, “Assuming a relatively complex account of consciousness, I find it intuitively hard to imagine how (e.g.) the 302 neurons of C. elegans could support cognitive algorithms which instantiate consciousness.” Muehlhauser does believe, however, that it is much more intuitive how the approximately 100,000 neurons of a crab might support cognitive algorithms which instantiate consciousness. ↩︎

6. Or philosophical theory. ↩︎

7. Although this could still be a surprisingly small number of neurons. ↩︎

8. Reptiles also have a cortex, but they do not have the six-layered neocortex found in mammals (Dugas-Ford, Rowell, & Ragsdale, 2012). ↩︎

9. See also Husband (2017) for an argument that there has been a historical bias in favour of the importance of cortex for consciousness. ↩︎

10. Shigeno, Andrews, Ponte, & Fiorito (2018) also show similarities between the octopus nervous system and the cerebral cortex. See also Fiorito et al. (2015), a report which contains a judgement that octopodes have cortex-like structures. ↩︎

11. The midbrain is evolutionarily more ancient than the cortex and is highly conserved among vertebrates. Because the midbrain is evolutionarily older, a greater number of different animals have a brain structure resembling a midbrain than a cortex. ↩︎

12. Also along these lines, Tye (2017) notes that decorticate rats act out purposeful behaviours that are indicative of consciousness. ↩︎

13. However, the wording here is a little murky and can be interpreted differently. While Barron & Klein (2016) cite the source as evidence in favour of the sufficiency of the midbrain, Allen-Hermanson (2016) views it as somewhat weak evidence. ↩︎

14. He adds that the telencephalon (the forebrain, containing the cortex) is important for elaborating the contents of consciousness. There is some debate about whether this means that consciousness can be supported solely by the midbrain (what is consciousness without its contents?). See Mallatt & Feinberg (2016, p. 8) and Klein & Barron (2016a, p. 4) for a discussion. ↩︎

15. Key (2016) also cites patients with analgetic thalamic syndrome as evidence that experience of pain is supported by cortical structures. Loss of cortical connectivity in these patients results in loss of pain sensation, despite subcortical structures remaining intact. ↩︎

16. In response to doubts about midbrain-sufficiency in humans, Mallatt & Feinberg (2016) have made an argument proposing that the seat of consciousness migrated from the midbrain to the cortex during mammalian evolution. This enables one to allow that the midbrain is not sufficient for consciousness in mammals, whilst maintaining that non-mammals with midbrain-like structures are conscious. ↩︎

17. It is not unusual in biology for different mechanisms to independently evolve to perform the same function: for instance, fishes accomplish the task of red-green colour discrimination using a different set of molecules to primates (Bowmaker, 1998). ↩︎

18. One might challenge this assertion with a concern about “hidden qualia”. See Appendix H of Muehlhauser (2017) for an explanation of this concern. ↩︎

19. A panpsychist view such as that advocated by Tononi, Boly, Massimini, & Koch (2016) would disagree with this statement, however. ↩︎

20. Many bivalves have sedentary or sessile adult lifestyles which do not involve navigation of the environment. However, this is not to suggest that bivalve behaviour can purely be explained by stimulus-response mechanisms. Some bivalves such as the European fingernail clam are more active (they climb on weeds to find the best feeding spot). It should also be noted that many young bivalves are mobile when finding a place to live. ↩︎

21. Honeybees and fruit flies likely have spatial memory and are likely to be able to navigate known and unknown paths (Rethink Priorities, 2019). See Barron & Klein (2016) for an argument that the insect brain has a specialised centre for processing spatial information and organising movement. ↩︎

22. In the sophisticated octopus nervous system more than half of the animal’s 500 million total neurons are distributed among the tentacles, which are each capable of somewhat independent behaviour (Sumbre, Gutfreund, Fiorito, Flash, & Hochner, 2001). This represents a quite radical distribution of information processing in the octopus nervous system. Despite this, a central executive still exists in the form of a central brain structure containing approximately 45 million neurons. ↩︎

23. One of the most prominent examples is “What insects can tell us about the origins of consciousness” (Barron & Klein, 2016) ↩︎

24. See Tomasik (2019) for a discussion of clock speeds in smaller animals. ↩︎