This invention relates to monitoring attention.

There are many occupations in which it is vital to maintain attention over a substantial time span. In many of these occupations, essentially boring activities are carried out for considerable periods of time and this can lead to loss of concentration on the part of the operator, sometimes with unfortunate or even fatal results.

Examples of occupations where continuous attention over a long time span is needed are public service vehicle driving, particularly coach driving on long haul routes, which is extremely monotonous and boring on motorways, piloting aircraft, and monitoring the operation of large installations such as chemical plants, oil refineries and power stations. Other analogous examples are coastguard and naval look-out duties, sentry duty, and the occupations of railway signalmen, security guards and night nursing staff. Particularly in the case of occupations which are carried out at night, there is a danger that the person concerned will have a fairly disrupted sleep pattern and may on occasion be required to carry out his job when his body wishes to be asleep. The danger of falling asleep on the specific occasion may well not even be apparent to the person concerned, much less to others, but the dangers of substantial loss of concentration before falling asleep or even falling asleep

itself hardly need to be stressed.

British Patent Specification 2015168 discloses an approach to the problem involving detecting a change in a physiological parameter, particularly a parameter of a driver's head. No details of how to do so are given.

U.S.S.R. Specifications 914360 and 994319 also deal with the subject. However, neither discloses a system which is both practical and efficient.

According generally to the present invention, a method of monitoring a person's alertness and detecting the onset of sleep or drowsiness includes the steps of contacting the person's skin with a pair of spaced electrodes, making a series of measurements of the electrical resistance between the electrodes, or continuously monitoring that resistance, and responding to an increase in the resistance by more than a predetermined amount. Preferably the electrodes are applied to a volar surface (defined herein as the palms of the hands, soles of the feet, and adjacent skin areas on wrist, ankle, fingers and toes on the same side of the hand or foot as the palm or sole respectively).

According to a further feature of the present invention there is provided apparatus for monitoring a person's concentration and warning of the onset of sleep or drowsiness which includes means for monitoring a physiological function or parameter which correlates with concentration, particularly one which correlates with impending sleep, and means for generating, if a change of more than a predetermined amount is detected, a suitable stimulus, wherein the physiological parameter is electrical skin resistance and wherein the apparatus includes means adapted to abut an extremity skin surface and means to emit a stimulus on detecting an increase in the electrical skin resistance thereof above a threshold amount.

In a particular embodiment, the invention provides

apparatus for monitoring concentration and detecting the onset of sleep or drowsiness which comprises an insulating carrier, a pair of spaced electrodes in the carrier, means for attaching the carrier to a person with the electrodes abutting the skin of the person at two spaced positions, means for checking the electrical resistance between the two electrodes, means for comparing the electrical resistance so measured at two spaced times, and means for generating a signal if the resistance has increased between the two spaced times by more than a predetermined threshold. The physiological parameter of electrical skin resistance is most preferably measured on a volar surface. The term volar surfaces as used herein means the palms of the hands and those parts of the fingers and thumbs and of the wrist on the same side of the hand as the palm, and the corresponding portions of the feet i.e. the soles and other downwardly facing surfaces. It is found that these give a particularly clear _, and highly sensitive indication, in terms of a change in electrical skin resistance, of the approach of sleep or of loss of concentration related thereto. The changes are conveniently detected electronically and accordingly can be simply continuously monitored to ensure immediate response should a dangerous situation arise.

The invention may conveniently be thought of in two portions, a first designed to provide a signal on the onset of sleep or loss of concentration and the second in terms of converting that signal into a stimulus, e.g. a warning, and feeding it back either to the person in question or to someone else, or using the stimulus to cause some other change, e.g. the initiation of a safety measure such as train braking. ^* It is of particular value

to use both feedback to the person in question and to generate stimuli elsewhere. For example in certain situations, it is desirable not merely for the person in question to know that he has just been in danger of falling asleep or losing concentration, but for others, for example co-workers, to know that as well.

A very wide range of annunciation devices may be used to impart a warning stimulus, varying from simple audible bleepers or sirens to flashing lights or even electrical arousal stimuli. For example, in the case of the measurement of skin resistance, which requires the application of two electrodes spaced apart to the skin, two further electrodes could be used e.g. in a common bandage but applied to the skin on the back of the hand (which is more sensitive) to feed a mild electrical shock to the person, which would provide an appropriate stimulus to wake the person up. In a particularly preferred embodiment of the present invention the amount of stimulus increases with time if the original enabling signal continues to be present. Thus a stimulus which restores the person in question to full concentration will cause a rapid change in the physiological parameter being detected and the stimulus will cease. On the other hand if the stimulus is insufficient to arouse the person in question, the stimulus then increases in intensity until the desired effect is achieved.

Alternatively, or in addition to an increase in stimulus, a secondary warning or stimulus may be activated after a preset delay, e.g. 10 seconds. Such secondary warning or stimulus may be a light, e.g. a flashing light or a further bleeper, either locally to the person in question or remote from him, or both. One specific secondary warning would be an automatic telephone message relayed to a distant point, e.g. to alert the next sentry, signalman or the like that his colleague had fallen asleep. In a further alternative, the secondary stimulus

may take the form of a signal which positively effects some further action, e.g. stopping a train being driven by someone who is drowsy.

The invention may be put into practice in many ways. One example is shown in the accompanying drawings in which:

Figure 1 shows schematically a simple attention monitor for use by long distance coach drivers.

Figure 2 is a circuit diagram for a skin resistance monitor i accordance with the invention, and Figure 3 is an alternative circuit. Referring to Figure 1, this shows a coach driver 1 sitting at the wheel 2 of a schematically indicated coach 3. In his left hand jacket pocket there is located a control box 4 connected electrically by a flexible lead 5 to a bandage 6 around the palm of the driver's left hand and by a second electrical lead 7 to a pair of headphones 8 worn by the driver. Within the bandage 6 is a pair of metal electrodes 9 of known type which come into contact with the skin of the palm of the driver and which are a suitable distance apart to measure the resistance between them. Control box 4 continuously monitors this resistance electronically.

If the driver loses concentration or begins to fall asleep, the electrical skin resistance of his volar surfaces rises, and this is detected by control box 4 as a rise in the resistance between the electrodes 9 set in bandage 6. A suitable electronic oscillator is thereby actuated and a high pitched buzz or other alarm sound transmitted via lead 7 to headphones 8. This startles the driver back into a state of concentration, whereupon the electrical skin resistance drops, this is detected by the control box and the noise in the headphones stops and remains stopped so long as the driver remains concentrating.

As shown the device is a simple portable device

having a bandage, earphones and a control box suitably battery operated and with a simple solid state circuit therein. The three components of this simple system are hardwired together but in certain circumstances this may be inconvenient and the signals may be transmitted at some stage in known fashion, e.g. via a radio or microwave signal, ultrasonically or via infra-red radiation. The control box and alarm system could, in such a case, be a built-in fixture powered by a mains or vehicular electric supply.

Referring now to Figure 2 this shows one example of a circuit suitable for incorporation in control box 4. For the sake of simplicity, the circuit shows a bleeper B, but this may of course be the headphones 8 connected to a box 4 via a suitable plug/socket connection.

The electrode assembly in the embodiment illustrated consists of a pair of skin-contacting electrodes 9 which are held, e.g. by a suitable elastic bandage device 10, against the palm of the hand of the person in question. he two electrodes 9 are connected via twin flex 11 to a jack plug 12 which is inserted into jack Jl shown on the left in Figure 2. The bandage 10 may have cooperating burr fastener pieces 13, 14 to enable it to be easily applied and removed. However, the electrode configuration and arrangements may vary widely, and electrodes can be incorporated into a glove, finger-ring, finger-stall or even in the knob on a "dead-man's handle" device as used in railway locomotives.

It is found in practice that the skin resistance measured on a volar skin surface varies very substantially from one person to another. It is therefore desirable that the circuit in box 4 takes this into account and according to a preferred feature of the invention the circuit should be self-calibrating, first effectively measuring the volar skin resistance and then registering a rise if one occurs.

It has furthermore become evident that there is a fairly long "settling in" period for simple metal electrodes in contact with the skin, so for some minutes after the electrodes have been placed against the volar surface, there is a gradual settling in and the electrical skin resistance appears to drop gradually, preferably the circuitry is designed to accommodate this phenomenon also. This may be done in one of two ways:

In a first approach, the circuitry is designed to 0 self-calibrate, i.e. to change its "alarm" threshold if the skin resistance drops more than a predetermined "settling in" threshold. Alternatively, because loss of concentration is unlikely in the first few minutes after placing the electrodes in contact with the skin, the 5 circuitry may include a delay mechanism which renders it inactive for e.g. 5 or 10 minutes after switching on, whereafter it will monitor the skin resistance and emit a stimulus if it rises by more than a predetermined threshold, as usual. 0 Finally, it is clearly desirable to measure the skin resistance in a smooth and reproducible fashion, and it is found in this connection that difficulties can be experienced if the applied voltage between the electrodes in contact with the skin exceeds about 2 volts. 5 It is also a preferred feature of any electronic system that self-calibrates that it gives an audible indication of when self-calibration is completed or gives an audible indication if the electrodes are either totally open circuit, i.e. an infinite resistance between them, or Q are short circuited one to another.

Referring now to Figure 2 in detail, the general circuit arrangement will be clear from a study of the diagram. As mentioned above, the skin contact electrodes are connected via a jack socket Jl shown on the left-hand 5 side of the diagram. Power is supplied via a jack socket J2 on the right-hand side of the diagram.

- O PI

Connected in series with the skin-contact electrodes is a constant current generator consisting of TRl and IC3. The value of the collector current of TRl is selected by the value of voltage applied to the non-inverting input of IC3. This value is in turn derived from a IM potentiometer RV3 which has a voltage applied across it from the output of a ten stage binary counter IC6 which drives an R/2R ladder network shown to the left of IC6 in the drawing. On powering up the circuit, IC6 resets to zero.

Driving a constant current between the electrodes produces a voltage at point A which is divided by a divider chain consisting of Rl and R2 to a lower voltage which is then applied to the inverting input of IC1 and the non-inverting input of IC2. Both IC1 and IC2 are FET input operation amplifiers (type Ca 3140) and the voltage at A is reduced by Rl and R2 to give a voltage at point B which is within the common mode range of both IC1 and IC2.

Initially, when the ON/OFF switch is turned ON, the current through TRl is zero so the voltage at point B is high giving a high output at the output of IC2 which in turn enables clock oscillator IC4/2. The output of this clock is fed to the ten stage binary counter IC6 which causes this counter to count up thus producing an increasing voltage from the ladder network which is applied across RV3 and accordingly an increasing voltage applied to the non-inverting input of IC3. This in turn causes an increasing collector current in TRl which causes the voltage at point B to fall. The inverting input of IC2 is connected to part of a resistor chain between the 9 volt positive rail and ground and the non-inverting input of ICl is likewise connected to that resistor chain, in this case via a 100K resistance.

When the voltage at point B drops to below that of the non-inverting input of C2, the output of IC2 goes negative and the clock IC4/2 stops. The circuit is now

calibrated. The voltage applied to the non-inverting input of IC2 is set by varying RV2 in the resistor chain mentioned above so that the voltage produced across the electrode at this stage is around 1.8 volts. If the resistance between the electrodes now decreases, this causes the output of IC2 to change, the clock to be enabled and the counter of IC6 will count up until the clock is disabled again. Thus, if the resistance between the electrodes drops during a settling in period, the circuit automatically re-calibrates.

Each time the circuit calibrates or re-calibrates, the output of IC2 goes negative and this is applied via IC4/3, IC4/4 and TR2 to produce a bleep on the bleeper, which is indicative of calibration being initially completed or a re-calibration having been carried out. During all of this, the output of IC1 remains the same since its threshold is set to a lower level compared with IC2.

If the resistance between the electrodes now increases, for example as a result of the onset of drowsiness, the voltage at point B drops, the calibration system cannot follow the change, and before too long the voltage applied to the inverting input of IC1 falls below that applied to the non-inverting input so causing the output of IC1 to go high. This output is fed to IC4/4 and this via TR2 drives the bleeper thus giving an alarm signal.

If the alarm alarms the wearer of the device, then the increase in attention thereby generated causes the electrical skin resistance to drop so the voltage at point B rises again and when this voltage, which is applied to the inverting input of IC1, corresponds to the voltage on the non-inverting input of ICl then the output of IC1 goes low again causing IC4/1 which is connected thereto via 1 Megohm resistor to produce a positive pulse which is transmitted to the reset port on IC6. This causes

re-calibration to occur. If it is not desired to re-calibrate, a switch may be used to ground the input of IC4/1.

The time delay network (1 Megohm resistor plus 4.7 F capacitor) between ICl and IC4/4 ensures that the bleeper sounds for a minimum time as soon as it is triggered.

If desired re-calibration can be effected at any time by pushing the CAL button which likewise puts a 9 volt positive pulse into the reset input of IC6.

As noted above it is very desirable to vary the alarm trigger level depending on the initial resistance between the electrodes. In the circuit shown this operates as follows: During calibration an increase in voltage is applied to the non-inverting input of IC3 until the current through TRl increases to the point where the voltage of point B drops to the calibration level. This may be chosen such that the voltage across the electrode at that stage is around 1.8 volts. The output of IC3 is accordingly inversely proportional to the resistance between the electrodes after these have settled in.

The input of IC3 is applied to an inverting amplifier IC5 so the higher the output of IC3, the lower of IC5.

As noted above, the level at which the alarm is triggered by ICl is set by varying RV1 and if the electrode resistance is about 2 Megohms, then the alarm level with the component values indicated may be set to about 20 per cent. This once done, the output of IC5 is adjusted by varying RV4 to give the same voltage at point C. At this point, no current flows into IC5.

However, as the resistance between the skin contacting electrodes decreases, this causes IC3 to drive TRl to increase the current in order to maintain calibration so the output of IC3 rises, the output of IC5

therefore falls and point B is progressively pulled lower in voltage. This inceases the threshold percentage increase in resistance necessary to trigger the bleeper. Because of the shape of the inverse functional 5 resistance produced by IC3/5, the percentage threshold increase needed to trigger would rise to unacceptable limits with a resistance between the electrodes of about 330K or below. To avoid this, the output of IC3 is fed to the line connecting IC5 to point C via three diodes and 0 resistor thus limiting the threshold increase below inter-electrode resistances of about 330K. Point D drops in voltage until diodes conduct.

The amount by which point B drops in voltage depends on the setting of RV5 which accordingly constitutes slope 15 control.

The circuit has the additional advantages of demonstrating instantly if it is working incorrectly. If, on switch on, jack 1 is open circuit, either because nothing is plugged in or because the electrodes are not 20 attached to the user, the voltage at point B remains high on switch on and the bleeper sounds continuously to indicate that the electrodes have not been properly connected or that there is an open circuit.

In the alternative case where the jack 1 is shorted, 25 the system starts to calibrate itself as previously mentioned but even when IC6 is counted through completely the highest voltage will be insufficient to pull down point B to the calibration level. The counter IC6 will accordingly step past its 10th output to its 11th output 2 _Q and as shown this is connected via a diode to the input of IC4/4 which in turn drives the bleeper via TR2.

A 10K resistor may be inserted in series with the bleeper by means of a loud/soft switch to vary the bleeper volume. _ ₅ Figure 3 shows an alternative circuit in which compared to Figure 2 IC5 and associated circuitry is

omitted. Compared with the circuit of Figure 2, the calibration range runs from 2 Megohms down to 45 Kilohms (Figure 2, 330 Kilohms), and the alarm trigger threshold is now at a constant percentage increase over the whole calibration range. The percentage increase required to trigger the alarm may be set anywhere from about 50% to about 320% by altering the setting of RV1. using the circuit with component values as in Figure 3, calibration at the higher volar resistances is very fast and no calibration pips are produced above an interelectrode resistance of about 1 Megohm. The large range of resistance thresholds enables the circuit to be matched to the user. "Low resistance" subjects may need a small threshold or a large one, depending on whether they exhibit only a small volar skin resistance increase when drowsy, or whether their skin resistance rises proportionately very rapidly.

OMPI *