Core symptoms of dementia are deficits of learning and memories and anti-dementia drugs are often called as "nootropics" or "cognitive enhancers". Since dementia patients recently often express their symptoms by themselves in media, we actually became to know that various kinds and levels (severe or mild) of dementia exist and our images for patient lost all intelligences have been changed. Dementia is classed into "Delirium, Dementia, and Amnestic and Other Cognitive Disorder" in DSM-IV. Eleven kinds of dementia such as Alzheimer's type, vascular, HIV disease, head trauma, Parkinson's disease, Pick's disease, and so on and dementia not otherwise specified are presented in it. Diagnostic criteria of dementia are expressions of the both 1) memory deficits and 2) cognitive deficits (aphasia, apraxia, agnosia, and executive dysfunction). Dementia is discriminated from amnesia by the presence or absence of the cognitive deficits and from delirium by continuation of the deficits. Anyway, because these disorders have "memory deficits (dysfunctions to learn novel information and to recall the information learned previously)" as the core symptoms, evaluations of learning and memories are essential to develop their therapeutic drugs. We must learn many things but learning in mice and rats that we can evaluate is not so many because we must detect it by their behaviors but not by words. Following 4 kinds of methods are mainly used on evaluations for learning in rodents: 1. learning of places, 2. learning of behaviors, 3. learning of stimuli, and 4. correct selection. For example, maze learning in articles 1, operant and avoidance learning in articles 2, taste aversion learning and fear conditioning in articles 3, and delayed matching to sample and drug discrimination in articles 4 can be listed. Correct selections in the delayed matching-to-sample test are evaluated by a choice in "places" and "behaviors" such as correct lever pressing or correct arm entry and various learning described above is often associated each other. There are a lot of actual methods to evaluate learning in rodents because various kinds of mazes, behaviors, stimuli and so on and their combinations are produced by researchers' originalities and ingenuities. The methods for learning in rodents are also categorized into evaluations of working memories (shortly valid in a trial) and reference memories (continuously valid between trials) based on memory durations but not subjects to be learned.

4-1 Maze Learning

Some mazes have several arms extended from a central platform on which rodents can move and others have different shapes. A maze consisted of multiple combinations of walls arranged as T-shape like a puzzle game exists but it is not so much used.

4-1-1 T Maze
4-1-2 Y Maze
4-1-3 Elevated Plus Maze
4-1-4 Radial 8-arms Maze
4-1-5 Morris Water Maze
4-1-6 Barnes Maze

4-2 Learning of Behaviors

Operant learning generally measured by lever pressing and learning of avoidance behaviors from electric shocks (passive avoidance and active avoidance) are listed. Especially, a passive avoidance test in mice is most often used as an easy screening method.

4-2-1 Operant Learning
4-2-2 Passive Avoidance Response
4-2-3 Active Avoidance Response

4-3 Learning of Stimuli

Electric shocks as negative reinforcing stimuli (punishment) and foods as positive reinforcing stimuli (rewards) are often used to motivate learning in rodents. Although the foods and the electric shocks are often used motivations in maze learning and learning of behaviors, escape from water in the water maze and escape from fear of open spaces in the elevated plus maze and the Barnes maze become motivations. Although the conditioned and unconditioned stimuli are presented in taste aversion and fear conditioning, aversive feeling by LiCl and punishment stimuli (electric shocks) are given as the unconditioned stimuli. Rodents learn that these punishments are given under certain condition.

4-3-1 Taste Aversion
4-3-2 Fear Conditioning

4-4 Learning of Correct Selection

Many of maze learning are learning of correct selection in a point that rodents select a correct arm from some arms. Delayed matching to sample or delayed nonmatching to sample in which selection of the same or the different stimulus after a sample stimulus is correct, respectively, are conducted by selecting an arm in a T maze and pressing either lever in a two-lever operant chamber. The lever pressing is usually used in the drug discrimination.

4-4-1 Delayed Matching to Sample
4-4-2 Drug Discrimination
4-4-3 Object Recognition

4-5 Closing

Because learning and memories in rodents are evaluated by their behaviors, we likely consider that no objective behaviors are dysfunction of the learning and memories. For example in the passive avoidance learning, no movement to avoid electric shocks is a parameter of memories but sedatives induce no movement and stimulants make easy to move in spite of memories. Because mice generally move well compared to rats, positive effects on learning and memories in rats likely detected well compared to mice. Because thus the behaviors in rodents are easily affected by other than learning and memories, we must carefully discuss the results and also methods less affected. For example, administration of lower dose of drugs at which general behaviors are not affected is possible. Moreover, if sedatives are evaluated in the active avoidance and stimulants in the passive avoidance, since behaviors judged as retention of memories are opposite to effects on the general behaviors by the drugs we can easily judge ameliorating effects of the learning and memories.

We must dementia model having dysfunction of learning and memories in rodents to evaluate anti-dementia effects of drugs. As a simple rodent model scopolamine-induced amnesia is widely used and as clinically mimicked models disorders of central neuronal systems in rats and knock-out or knock-in mice deficit or expressing specific genes are also studied. Additionally, various models are used such as a senescence accelerated mouse P8 strain developing an innate dysfunction of the learning and memories. Dysfunction in these models is generally not specific to learning and memories, and we must select the dementia model and evaluate the study results considering psychic and physical effects by the dysfunction in the models.