Humans are a very popular species in memory research. Over the years, far more memory experiments have been published on humans than on "monkeys, or on all invertebrate species combined. Only rodents still keep an edge in popularity in labs that specialize in the neurobiology of memory, but this could change, with the widespread availability of "functional neuroimaging. And yet, human beings almost always identify themselves with the side of the experimenter, even in situations in which they themselves suffer the fate of the experimental "subject, be it involuntarily in evolution or disease, or voluntarily in the lab. This raises some additional doubts on whether Homo (from 'earth' in Latin, a reminder of old myths) is indeed always sapiens (from 'to be wise = to taste' in Latin, another reminder of a biblical story).
There used to be additional Homo (Wood and Collard 1999). At least with one of these, neanderthalen-sis, we shared this planet for a while. It is questionable whether we remember the experience, a sort of "infantile sapiens amnesia, although it is still possible that some reminiscences do linger in the obscure legends of our "collective memory. It is generally assumed that sapiens emerged 1-2 x 105 years ago, probably in Africa, and made the first massive out-of-Africa exodus 0.6-1 x 105 years ago (Quintana-Murci et al. 1999).1 Assuming 20 years per generation, it means that the distance between us and the first true human may be only 5000 generations, not an astronomic number (try to imagine several thousand individuals holding hands; this is it). What was the brain power of the lost hominids species? And which cognitive capabilities, memory included, enabled us to win over the Neanderthal, or at least to linger longer? Attempts to unearth the answers combine prehistoric archaeology with anatomy, molecular biology, and common sense (Wilkins and Wakefield 1995; Wood 1996; Wood and Collard 1999; Yamei et al. 2000). Our current ideas on the memory "capacity of our ancestors are mostly speculations. The scientific community will no doubt be delighted if one day a method is devised to determine the problem-solving ability or "working memory "capacity in early hominids (cloning from hominid DNA could yield big surprises, but is unlikely to occur). However, at this point in time, the only clues to human memory emerge from what we get from experiments on living humans.
The study of memory in general had probably started with human subjects. First it involved sporadic observations intermingled with philosophical speculations and pedagogical generalizations (e.g. Sorabji 1972). These were followed by systematic introspection, in both its pre-scientific (Quintillian 1C ad; Augustine 400) and scientific versions (e.g. Wundt in Germany, Titchener in the USA, Boring 1950a). Objective "methods for quantifying memory were first documented only a bit more than a century ago (Ebbinghaus 1885). Darwinism, combined with practical considerations and the need to use methods that would not have been tolerated on humans, have provided the background and incentive for the study of memory in experimental animals (e.g. Boakes 1984). But even when animals became preferred subjects in the study of learning, a prevalent notion was that they are primarily convenient "models for understanding humans and the general 'laws' of mind: 'Most of the formal underlying laws of intelligence ... can still be studied in rats ... more easily than in men... (rats) do not go on binges the night before one has planned the experiment... they avoid politics, economics and papers on psychology (Tolman 1945; "rat, "simple system).
Let us list briefly the pros and cons of conducting memory research on human subjects. First, here is a short list of selected pros:
1. Only humans have a human brain. Whatever we find about the "engram in experimental animals, even in tasks in which these animals excel (e.g. "maze), or in animal models of human pathology (e.g. "amnesia), must still be adapted to and verified in the human brain.
2. The highest forms of learning, involving language, reasoning, and imagery, as well as intricate emotional experiences and the memory of the self (Conway and Pleydell-Pearce 2000), can only be studied in humans. Attempts to identify rudiments of these capabilities in non-human species are far from yielding satisfactory alternatives to human subjects ("anthropomorphism, "declarative memory, "episodic memory, "monkey, "Ockham's razor).
3. Humans (usually, admittedly not always) can follow instructions quickly and efficiently (the reader is cordially invited to take a breath and imagine, eyes closed, what is it like to be a lonely, frightened, perplexed "mouse in its first encounter with a bizarre problem box).
4. Only humans can report their experience verbally, in detail and in response to specific questions.
5. Human volunteers take care of themselves. No need for animal rooms, costly maintenance, even not quality-time according to NIH guidance. Unless they are a real nuisance, human subjects come and go as requested, and the experimenter can simply forget about them in between experiments.
And now here is a selection of cons, for a balance:
1. Invasive methods are out of question. This is probably the critical disadvantage of working with humans. We clearly cannot test a hypothesis, or replicate an amnestic pathology (e.g. Scoville and Milner 1957), by inducing brain lesion in a human subject2. Only observations in the course of mandatory medical treatment, e.g. surgery, are allowed. New functional neuroimaging methods provide alternatives for certain invasive experimental designs (such as the recording of neuronal activity from the insides of the brain) but not for others (lesions). Gone are even the days when psychologists were allowed to frighten babies to advance their own career (Watson and Rayner 1920; Jones 1930). Intentional "neurogenetic manipulations are also a no-go.
2. In some protocols, overenthusiastic human subjects are particularly prone to 'demand characteristics' and other sorts of "biases that could undermine the validity of the results. This, however, may not be unique to humans ("Clever Hans).
3. You do not have to be a scientist to discover that humans are not easy to work with. In most places it is difficult to get a sufficient number of human subjects for lengthy experimentations, unless, probably, you are the dean or run a prison. And in most of the cases the subjects available are heterogeneous in their biology and personal history. Further, most human subjects tend to shy away from though, or boring, tasks. They may start the experiment and then disappear. Animals have no choice, although even monkeys revolt from time to time. Animals also do not request payment for their participation in an experiment and surely will not request a raise. Also, animals are not required to sign a letter of consent, and will not sue the experimenter even if they become confident that their neurosis is due to their experience as subjects in a demanding experiment. It is not surprising, therefore, to discover occasionally that the acronyms of subjects in human psychophysics or neuroimag-ing papers are those of the names of the authors themselves.
Over the years, a large variety of similar tasks have been used in the investigation of both human and animal memory. On the one hand, sheep were trained to recognize faces (Kendrick and Baldwin 1987), horses convinced to read prose (Pfungst 1911, they cheated), and monkeys to master math (Kawai and Matsuzawa 2000). On the other, human subjects were expected to salivate like Pavlovian dogs (Lashley 1916), run in mazes like rats (Woodsworth and Schlosenberg 1954), and fear unexpected noises like rabbits (Watson and Rayner 1920). In general, in memory research it is best to choose a task that fits the species and permits it to disclose its full potential. As is the case with other species, humans are better in some tasks than in others, for example, in problem solving in a variety of environments, or in "planning. But we are not the ultimate memory machine. Rodents are superior in labyrinths, dogs in olfactory discriminations, bats in spatial localizations, and chicks in telling one individual hen from another ("imprinting). Even in some human-specific memory tasks we perform rather miserably, e.g. assigning names to faces (McWeeny et al. 1987); this is probably due to the fairly recent appearance of human names in evolution. To our favour it should be added that we are unique in trying to overcome by technology the limitations of our innate, "a priori capabilities, and improve our memory ("mnemonics, "nootropics). We also create new artificial worlds to which our memory systems must rapidly adapt. Virtual reality (e.g. Maguire et al. 1998) is but an example. The Web is another. Such new environments impose great demands on "working memory, the ability to categorize, and the "plasticity of cognitive "skills. They also open the possibility of intimate interfacing between human brains and electronic systems. In the future, some engrams may actually be engraved ("metaphor) in neurosilicon hybrids.
Selected associations: Amnesia, Anthropomorphism, Enigma, Monkey, Real-life memory
'Estimates of the structure and timetable of hominid lineages are characterized by substantial uncertainties and large standard deviations. The discovery of even a single new type of skeleton, or of a new site with hominid artefacts, could shift these estimates remarkably. See, for example, Leakey et al. (2001); also Humans on the move, Science, 291:1721-1753 (2001).
2Transient functional lesious using transcranial magnetic stimulation (TMS, see *amnesia) are explored but their future as acceptable routine experimental procedure is doubtful.
Was this article helpful?