SPINAL THERAPY APPARATUS

Abstract

A spinal therapy apparatus includes one or more manipulating assemblies. In some examples a first manipulating member is arranged to engage a spine at a first vertebral area between a spinous process and a transverse process on a first side of a spine and a second manipulating member is arranged to engage the spine at a second vertebral area between a spinous process and a transverse process on the first side of the spine. The first and second manipulating members are coupled together at their proximal ends to permit the first and second manipulating members to be driven simultaneously and the coupling permits motion of the first and second manipulating members relative to one another in an axial direction. In other examples a drive system is adapted to drive first and second manipulating members simultaneously and independently. The disclosure also relates to a spinal therapy bed including such manipulating assemblies and methods of using and training spinal therapy devices.

Claims

1-91. (canceled)

92. A method of control for a spinal therapy bed, the spinal therapy bed comprising: a plurality of manipulation members spaced along the length of a user's spine and adjustable to the curvature of an individual user's spine and arranged to apply pressure from either side of the spine at multiple points along the spine, each manipulating member arranged to engage a user's spine at a respective vertebral area between a spinous process and a transverse process arranged to be driven from a proximal end and having a portion for engaging the spine at a distal end; and a drive system; wherein the method comprises: controlling the drive system to actuate the manipulation members to apply pressure to their respective vertebral areas, wherein a plurality of the manipulation members are driven simultaneously and independently, wherein the spinal therapy bed further comprises sensors for monitoring the behavior of each manipulating member and wherein the method includes: receiving measurements from the sensors, the measurements comprising at least a measurement indicative of a force applied by at least one manipulating member over a cycle in which the manipulating member moves towards and away from its respective vertebral area and returns to its starting position; and adaptively adjusting the force applied by each manipulating member in response to the received measurements.

93. The method of claim 92 wherein a measurement indicative of a force applied is derived based on the current supplied to an actuator.

94. The method of claim 92 wherein the spinal therapy bed further comprises sensors for monitoring the distance travelled by each manipulating member and wherein the method includes: receiving measurements from the sensors, the measurements comprising at least a measurement indicative of a distance travelled by at least one manipulating member over a cycle in which the manipulating member applies a time varying force to its respective vertebral area; and adaptively adjusting the motion of each manipulating member in response to the received measurements.

95. The method of claim 92 wherein applying the pressure comprises changing a control variable: receiving a profile of the control variable over the cycle; identifying a threshold value of the control variable from the profile, wherein the threshold value is indicative of a time at which a statistically significant change in the control variable is determined; and controlling the drive system based on the threshold value.

96. The method of claim 95 wherein the control variable is selected from current, power or voltage drawn by an actuator.

97. The method of claim 95 wherein the threshold value of the control variable is updated throughout a treatment regime.

98. The method of claim 95 wherein the threshold value of the control variable provides an indication of the relative firmness of the vertebral areas; and a treatment regime is adaptively developed based on the indication of relative firmness of the vertebral areas.

99. The method of claim 97, wherein the treatment regime includes: controlling or changing distances for each manipulating member to extend during treatment; controlling or changing forces for each manipulating member to exert during treatment; and/or controlling or changing relative start points for each manipulating member.

100. The method of claim 92, wherein the independent driving includes: rippling motion with a constant phase between adjacent manipulating members; adjacent manipulating members being driven in antiphase; and/or different motions in cervical, thoracic and lumbar portions of a user's spine.

101. The method of claim 92, wherein the spinal therapy bed includes an adjustable leg support and the method includes adjusting the height of the adjustable leg support at specific timings during a treatment process.

102. The method of claim 92 wherein the drive system is fitted with a cut out operable to stop the driving system from driving one or more manipulating members if an actuator draws more than a threshold value of current, voltage or power.

103. A spinal therapy apparatus arranged to deliver a spinal therapy comprising: a plurality of manipulation members spaced along the length of a user's spine and adjustable to the curvature of an individual user's spine and arranged to apply pressure from either side of the spine at multiple points along the spine, each manipulating member arranged to engage a user's spine at a respective vertebral area between a spinous process and a transverse process arranged to be driven from a proximal end and having a portion for engaging the spine at a distal end; a drive system; a controller for controlling the drive system to actuate the manipulation members to apply pressure to their respective vertebral areas, wherein a plurality of the manipulation members are driven simultaneously and independently, a force sensing arrangement for determining at least a measurement indicative of a force applied by at least one manipulating member over a cycle in which the manipulating member moves towards and away from its respective vertebral area and returns to its starting position; and wherein the controller in configured to adaptively adjust the force applied by each manipulating member in response to the received measurements.

104. A spinal therapy apparatus according to claim 103 including control logic arranged to perform a method according to claim 92.

105. A method of control for a spinal therapy device of claim 103, the method comprising: determining control signals for controlling a plurality of manipulating members of the spinal therapy device based on a stored treatment profile, wherein the plurality of manipulating members are arranged to engage a user's spine at a respective vertebral area between a spinous process and a transverse process; applying the control signals to drive the manipulating members in at least one cycle to implement a treatment regime; receiving feedback signals from a plurality of sensors indicative of at least one of pressure and movement applied on the user's spine by the plurality of manipulating members during the at least one cycle; and determining adjusted control signals based on the control signals of the stored treatment profile and the received feedback signals to implement an updated treatment regime.

106. The method of claim 105 wherein the adjusted control signals are determined using machine learning algorithms.

107. The method of claim 105 further comprising applying the adjusted control signals in a subsequent cycle, optionally further comprising updating the treatment regime based on the adjusted control signals.

108. The method of claim 105 further comprising storing a new treatment profile based on the updated treatment regime.

109. The method of claim 105 wherein the stored treatment profile is selected from a library of treatment profiles, optionally stored at a central server, optionally wherein the library comprises treatment profiles selectable from one or more of the following: a general treatment profile; and/or a personal treatment profile developed for a user based at least in part on feedback signals from the user.

110. The method of claim 105, the method further comprising generating a general treatment profile for a spinal therapy device, the method further comprising: receiving at a first plurality of spinal therapy devices, a first treatment profile from a library comprising a plurality of treatment profiles; performing, the method of controlling the spinal therapy device of any one of claims 14-18 using the first treatment profile as the stored treatment profile; receiving by a central unit, feedback signals from the plurality of spinal therapy devices; processing by the central unit, the feedback signals to determine a general treatment profile to be used in a subsequent treatment; and storing the general treatment profile in the library.

111. The method of claim 110 further comprising; generating a plurality of general treatment profiles based on performing the method of claim 110 on multiple treatment profiles in the library.

Description

[0087] Aspects of the invention will now be described in detail with reference, by way of illustration only, to the accompanying Figures, in which:

[0088] FIGS. 1 and 2 illustrate one of a plurality of manipulating assemblies that form part of a spinal therapy device of an embodiment of the present invention.

[0089] FIGS. 3 and 4 illustrate an embodiment of the present invention.

[0090] FIG. 5 illustrates the movement of a Bowden cable bracket driven by a reciprocating actuator.

[0091] FIG. 6 illustrates a section of an embodiment of the present invention.

[0092] FIGS. 7 and 8 illustrate a massage table frame comprising an embodiment of the present invention.

[0093] FIGS. 9 and 10 illustrate elements of a massage table for an embodiment of the present invention.

[0094] FIG. 11 illustrates a cross section of the spinal therapy bed highlighting an angle relevant to user comfort in certain embodiments.

[0095] FIG. 12 illustrates a graph showing current applied to manipulating members against time.

[0096] The present invention comprises the application of one or more forces or pressures applied by a plurality of manipulating members to the soft tissue located adjacent each side of the vertebral column in the vertebral area between the spinous and transverse processes. The one or more forces are applied partially towards the base of the area between the spinous and transverse processes and partially towards a second side of the vertebral column opposite the first side such that a substantial length of vertebral column is rotated or rocked by the action of said forces on a plurality of vertebral areas between the spinous and transverse processes, wherein the plurality of vertebral areas may experience different applied forces caused by an imbalance of tension in said vertebral area.

[0097] To create movement of the spine, the area of the spine between the spinous and transverse processes should be engaged. This can be achieved by coordination of manipulating members such that they are configured to act as a pair, contacting both sides of a vertebral area about the spine. A substantial amount of force may be required in persons with stiff back problems, or indeed varying forces may be desired at locations with varying levels of tension along the length of the spine.

[0098] The aim of the apparatus is to create movement of the spine to cause the vertebral junctions to loosen up and relieve tension. The vertebrae of the spinal column may be required to be independently articulated to accommodate for the differences in physical stiffness or tension in each individual junction along the length of the spine.

[0099] Loosening of the vertebral junctions can be achieved by articulating a plurality of manipulating members at each vertebrae, one member on each side of the spinal column, such that a set of manipulating members may be independently operated to achieve a personalised treatment profile that can be adapted to treat variations both along a spine and for any spine.

[0100] The movements are applied uniformly, gradually, firmly and over prolonged periods of time. A suitable frequency of movements lies in the range of about 6 to 10 movements per minute, although other frequencies may also impart benefits.

[0101] Although a reasonably substantial amount of force is often required, caution must be exercised not to apply excessive force in local points so as to not cause bruises, pain, excessive discomfort or to further inflame an existing injury.

[0102] Individually addressing each articulating or manipulating member of a set of members that are required to manipulate the full length of the spine requires intensive energy and resource use to operate the members at each desired intensity for the particular location of the spine that the member is addressing.

[0103] The present invention provides a device for manipulating the vertebrae using a plurality of manipulating members that can be individually addressed whilst coupling adjacent manipulating members such that they may be addressed as a coupled pair but have a degree of relative movement, using a single motor to drive the coupled pair, rather than being individually addressed. Advantageously, this reduces the energy used and audible noise in use, as well as streamlining the device, whilst maintaining a fully individualised massage treatment experience for the user.

[0104] The present invention provides a number of advantageous features that allow spinal therapy to be performed. These features are provided by manipulating members in a manipulating assembly comprising a means for coupling the members such that relative movement between them can be achieved such that two manipulating members can be addressed at a proximal end using a single driving source to drive the coupled members whilst providing personalised treatment at a distal end; a spinal therapy bed configured to address particular regions of the spine by virtue of the spacing of manipulating assemblies; a “floating” motor that provides control of the manipulating members and ensures that the force exerted on the spine does not exceed a threshold that may cause pain or damage to the user and which does not transmit vibrations to the chassis, so is quieter and more comfortable; and a method of controlling the spinal therapy bed that provides a personalised treatment.

[0105] Referring to FIG. 1 and FIG. 2, an apparatus according to an embodiment of the present invention is described herein.

[0106] A manipulating assembly 100 comprises an assembly including a first manipulating member 10, a second manipulating member 15, a third manipulating member 20, and a fourth manipulating member 25, mounted together, each of which may be configured to engage the spine at a different vertebral area and may each be independently addressable and controllable.

[0107] Each of the members in the assembly comprises an elongate rod 35, which has a cap 70. The rod can be made from a material that is suitably strong and resistant to deformation yet reasonably lightweight, such as stainless steel. The elongate rod may have a diameter between 10 mm and 15 mm, more preferably wherein the diameter is about 12 mm. The tip of the elongate rod is configured to have similar shape and hardness to a fingertip and has a diameter of between 5 mm and 10 mm, more preferably about 8 mm. The cap may be rubber or another material that is pliable and provides a cushioned interface between the spine and the elongate rod 35. The cap may have a diameter of 12.5 mm and a thickness of 2.25 mm. The pliable pad may have a hardness of 60 or less, as measured on the Shore “A” scale. The elongate rod 35 moves linearly along an axis defined along the longer side of the rod, along its centre, the movement being controlled by an actuator comprising a motor 65, a coupling 60 and a lead screw 55. A lead screw 55 translates the rotational motion of the motor 65 to linear movement, which moves the manipulating members to which it is coupled up and down along the axis of the elongate rod 35 such that the manipulating members penetrate the vertebral area at an angle of 50 degrees to the horizontal defined by the orientation of the supporting head 30. It will be appreciated that other actuating means including pneumatic, hydraulic, spring decoupled designs, and Bowden cable driven designs may be used to drive the coupled manipulating members, some of which are described in detail below. Indeed any suitable actuator may be used, many of which will be familiar to skilled workers in the field

[0108] Distances between the manipulating members in an assembly vary depending on the spinal region which they are designed to treat. A distance between the manipulating members shown in FIGS. 1 and 2 that address areas across the spine from each other is between 50 to 80 mm and depends on the area of the spine which is being treated. In the cervical region, the distance is between 50 and 70 mm. In the thoracic and lumbar regions, the distance is between 65 and 80 mm. The distance is measured from a centre of the cap 70 on a first manipulating member to the centre of the cap 70 of the manipulating member on the second side of the spine. A distance measured centre-to-centre between the caps of the fingers in a direction along the spine is typically a similar distance in all of the regions, and is about 35 mm. The first and second manipulating members may be spaced apart by a first distance in a direction along the spine, which is transverse to their axial direction of motion. The third and fourth members may also be spaced apart by the same first distance in the same direction along the spine as the first and second members. The first distance may be calculated centre-to-centre between the manipulating members, and may be between 30 mm and 40 mm, more preferably wherein the spacing is 35 mm. In adjacent assemblies in a same group, the distance between manipulating members measured centre-to-centre along the length of the spine between the caps is the same, even between non-coupled members.

[0109] The drive system may comprise a drive assembly for each respective manipulating assembly, each drive assembly comprising: a first actuator for driving a first manipulating member pair; a second actuator for driving a second manipulating member pair; wherein the control system simultaneously activates the first and second actuators to drive respectively the first and second manipulating member pairs. In some embodiments, the drive system can be configured to drive the spinal therapy device comprising a number of drive assemblies according to a predetermined program. For example, an actuator may be provided by a motor and lead screw arrangement. Other driving mechanisms including pneumatics or hydraulics may also be used in some examples.

[0110] A spring 40 is placed between the pivoting member 45 and the support head 30. The spring allows the assembly to smoothly return to a start or rest position. If the assembly is forward driven, as in FIGS. 1 and 2, and the spring has a high spring constant, the spring 40 could also be used to provide a certain resistance against the members being pushed too hard or too fast into the back of the user, i.e. to provide a degree of protection to the user from overextension of the manipulating members.

[0111] In one example of a full cycle of the motor drive, the elongate rod 35 is linearly driven forwards from a resting position by a displacement of about 20 mm in the direction of the user's back then returns backwards to the resting position, aided by the spring 40.

[0112] The frequency of the linear movements of the elongate rod caused by the configuration of the motor 65, the coupling 60 and the lead screw may be between 3 Hz and 6 Hz, more preferably 4 Hz. Manipulation of the tissue of a user resting on the manipulating members preferentially causes relaxation and loosens the vertebral junctions.

[0113] The first member 10 and the second member 15 are coupled together at their respective proximal ends by virtue of a pivoting member 45 and are driven as a couple in the same direction and at the same time. The members are arranged such that they address a first side of the spine at adjacent vertebral areas. The pivoting member 45 is a swivel bracket, as shown in FIGS. 1 and 2, arranged to pivot about a centre-point between the first and second members. The pivoting member 45 is supported by a bracket 50, which drives the assembly in response to the actuator and provides a backstop that stops the swivel bracket from pivoting too far. The first and second members are spaced apart by a distance of about 33.4 mm.

[0114] A stopper 70 provides a soft interface between the elongate rod 35 and the pivoting member 45. This allows for a degree of rocking of the members to distribute load evenly on the user's back for individualised treatment and to protect against jolting movements which may be uncomfortable or cause injury or damage the device. The stoppers 43 shown in FIG. 1 comprise rubber grommets, but other similar configurations may be used to provide this effect.

[0115] The pivoting member 45, driven by the actuator, is pivoted by a difference in resistance of the manipulating member against the user, caused by a difference in tension or stiffness inherent to the tissue around vertebral junctions. If there is a resulting difference in the force applied to the elongate rods 35 of the first and second members by the interaction between the spine and the manipulating members, a differing range of motion of the adjacent first and second members may be experienced. For example, if the first member 10 is manipulating an area that is particularly stiff or tense, it may have a restricted range of movement compared to a second member 15 that manipulates an area that is less stiff or tense. This will cause the pivoting member 45 to pivot about the central point between the members, such that the range of motion of the second member 15, in synchronicity with the movement of the members by the actuator, is greater than that of the first member 10. The pivoting may be resisted or limited by elastic materials, springs, frictional bearings, etc. to alter the response of the pivoting member 45 to tissue stiffness differentials. This can help ensure that the relative axial motion of a coupled pair of manipulating members 10, 15 is appropriate in response to particular tissue stiffness differentials.

[0116] A third manipulating member 20 is positioned on a second side of a vertebral area to the first manipulating member 10, such that they form a pair about the same vertebral area or around a single vertebra. This pair may be advantageously controlled to manipulate the same vertebral area from each side of the spine in a personalised manner, which may be synchronous or asynchronous.

[0117] A third manipulating member 20 and a fourth manipulating member 25 are coupled in the same way as the first member 10 and the second member 15 and are arranged to address a second side of the spine across from the first side.

[0118] An assembly of the four members addresses both sides of the spine at two vertebral areas, wherein the first and second members are arranged to engage one side and the third and fourth members are arranged to engage the other side; and wherein the first and third members address a first vertebral area and the second and fourth members address a second vertebral area.

[0119] The first, second, third and fourth members are assembled in a manipulating assembly 100 as shown in FIGS. 1 and 2. The members are supported and grouped together by a supporting head 30, which forms part of a Y-shaped bracket, comprising a support shaft 85 and the supporting head 30. The angle at which the members manipulate the spine is defined by the Y-shaped bracket, which can be in the range of 40 degrees and 60 degrees preferably where the angle of treatment is 50 degrees relative to the horizontal axis, or equivalently, around 40 degrees to the vertical axis and/or the leg of the Y-shaped bracket.

[0120] By virtue of the support head 30, and Y-shaped bracket in general, the number of parts of the assembly is reduced, which beneficially reduces manufacturing time and cost. It also provides a greater surface area of a platform for load bearing of the user or patient.

[0121] The supporting head 30 provides a greater surface area for taking the load of the user with respect to the manipulating members, helping to prolong the lifetime of the spinal therapy device and providing further comfort to the user. The members are able to move independently of the supporting head 30 and each other. They are arranged in a square or rectangular configuration and are evenly distributed about the central point of the Y-shaped bracket.

[0122] The manipulating members interact with the supporting head 30 which provides positional support but allows the members to move with a degree of freedom in the axis along the centre of the elongate rod 35, for example as shown in the Figures each manipulating member is arranged to slide through a respective aperture in an arm of the Y-shaped bracket.

[0123] The support shaft 85 onto which the support head 30 is mounted is provisioned with a retraction spring 75 and/or a plurality of locking slots 80. The retraction spring 75 provides greater flexibility of the assembly positioning on the support shaft 85. The plurality of locking slots 80 allow the manipulating assembly to be arranged at a personalised height. Adjacent assemblies may be provided at different heights to adjust for the curvature of the spine. This ensures the manipulating members maintain contact with the vertebral areas of the spine along the full length of the spine to improve the effect of the massaging and may also increase the comfort of the user.

[0124] The locking slots 80 can have a number of increments for individualised user experience. Configurations with greater or fewer numbers of increments are also possible. The slots are designed to be angled so that the locking plate that connects with the slot pushes inwards and upwards to maintain a connection with the user's back during locking. The edges of the locking slots are angled to allow smooth and controlled entry of height setting pins to avoid sharp or jolting movements.

[0125] For different regions of the spine, the support shaft 85 has different lengths. Where the manipulation assembly interacts with the neck, for example in the cervical region of the spine, the support shaft 85 has a length longer than the length of the support shaft 85 along the upper and lower back, for example the thoracic and lumbar regions. For example, the support shaft in the neck region may have a length of 217.5 mm and the length of the support shaft 85 in the upper and lower back regions may be 177.5 mm. The diameter of the support shaft 85 may be 25.1 mm, and it may have a D-shaped cross-sectional area that is shaped to prevent rotation of the support head 30. It will be appreciated that other dimensions and configurations may be possible.

[0126] The support shaft 85 may also provide a means for rotation of the manipulating assembly to provide further personalisation for users with spinal curvature.

[0127] In other configurations, not shown in the Figures, the swivel bracket that forms the pivoting member 45 may alternatively be a resiliently deformable block or a ball and socket joint for improving individualised treatment.

[0128] It is further possible to combine the manipulating members in assemblies having more than two pairs of manipulating members, for example three, four or six pairs.

[0129] The elongate rod 35 has a circular cross-section, however, it may be configured to be D-shaped, at least slightly, to prevent it from rotating about the axis along which it moves.

[0130] Other configurations of the manipulation assembly arranged by the support head 30 may be possible, such as assemblies configured to constrain more than two manipulating members on either side of the assembly; or more than a total of four members, or two pairs of members.

[0131] In some embodiments, the first, second, third and fourth members may be arranged such that their respective distal ends form a rectangle configuration, which may include a square configuration. Other examples include trapezoidal, rhomboidal and parallelogram arrangements, depending on the specific treatment required by a user. In some examples the arrangement of the tips can advantageously be combined with the relative timing of driving different manipulating members, so as to manipulate vertebrae in a particular manner, for example rocking, twisting and/or dragging motions may be achieved with particular combinations of distal end arrangements and the timing of actuation of each manipulating member relative to one another. Such motions may beneficially massage the tissue to release tension.

[0132] Although it will be understood that in general, coupling manipulating members and driving them as a pair is advantageous, manipulating members may be driven individually and independently, for example each manipulating member may be driven by an actuator for that manipulating member. Manipulating the manipulating members individually can provide a personalised treatment profile.

[0133] The drive system can be adapted to drive manipulating members on the first side of the spine out of phase with manipulating members on the second side of the spine, or to drive the manipulating members on the first side of the spine and manipulating members on the second side of the spine in antiphase with one another. In other embodiments, the manipulating members may be driven in phase or both in anti-phase and in phase throughout a treatment profile, for varying durations, which may be more effective than driving the members consistently at the same phase. The drive system can drive the manipulating members with a constant cycle time or with a varying cycle time between the different members and can be configured to control assemblies individually and/or in combination with other assemblies, for example, according to a pre-determined program of the spinal therapy device.

[0134] FIGS. 3 and 4 illustrate an example embodiment of an array of manipulating assemblies arranged such that they are positioned along the length of a spine in a spinal therapy bed configuration.

[0135] A manipulating assembly 100 is mounted on a chassis 90. Manipulating assemblies are arranged on the chassis 90 in groups that correspond to different areas of the spine, namely the cervical, thoracic and lumbar vertebral regions.

[0136] The cervical region, at the top of the spine near the neck, is the smallest of the regions; the thoracic region is the middle and largest region; and the lumbar region is at the lower end of the spine adjacent the pelvis. Each of the regions has differing basic vertebral structures, such that the way in which they are to be manipulated varies from area to area. Manipulating assemblies adjacent one another form a group that corresponds to one of the regions of the spine.

[0137] The cervical region of the spinal therapy device comprises one manipulating assembly, each manipulating assembly comprising four manipulating members; the thoracic region comprises three manipulating assemblies and the lumbar region comprises two manipulating assemblies as shown in FIGS. 3 and 4.

[0138] The spacing of the groups is greater than the spacing of the manipulating members in the assembly. Preferably, the spacing between groups is 105 mm. The spacing of manipulating members in an assembly in a single group is less than this and is about 35 mm. The spacing between adjacent members in a group in a direction along the spine is maintained at a constant value such that the tip-to-tip distance between all of the manipulating members is approximately the same within a group and an assembly.

[0139] The assemblies that are arranged to treat the vertebrae in the cervical region have a spacing of manipulating members that is 65 mm, and the spacing of members in the thoracic and lumbar regions is 70 mm. The cervical vertebrae are smaller and thus closer together, which is reflected in the spacing. By mirroring the spacing of the assemblies with the spacing of the vertebrae in each region of the spine, a more accurate treatment can be given and a greater overall result of loosening the vertebral junctions may be achieved.

[0140] FIG. 5 illustrates the motion induced by the actuator, which is driven by the motor 65. The motor 65 drives the lead screw 55, which in turn provides movement to a Bowden cable bracket. The lead screw may be, for example, a T8×2 mm lead. In other examples, the reciprocating motion may be provided by other means, such as a rack and pinion gearing system. The bracket can move at 20 mm/s. The assembly is capable of moving 20 mm in either direction about a centre point and may be fitted with sensors that can prevent over driving the bracket. Bowden cables provide flexibility of force transfer amongst the apparatus, and require decoupled motion, which means that there is no chance for a direct drive from the motor to push the members straight into the back, which can be uncomfortable for a user if the force is too large. Bowden cables can also be joined between a pair of manipulating members on opposing sides of the spine such that they can move in synchronised anti-phase. If the Bowden cable's radius of curvature is too small, this can create unwanted friction. This can be improved by stepping the motor-to-member-cable connection, for example by driving the fourth member using the first motor.

[0141] FIG. 6 illustrates a Y-shaped module or bracket, which is one embodiment of the support head 30. The Y-shaped module advantageously provides a platform to support four manipulating members on one retraction pole or support shaft 85. This beneficially reduces the number of parts required for construction of the spinal therapy bed and simplifies the general architecture.

[0142] FIGS. 7, 8 and 9 illustrate the spinal therapy device in part of a massage bed assembly. The assembly comprises a bed structure 200, where the user or patient lies down face up on their back, and a footrest portion 220 that supports the lower body. The portion of the bed onto which the user lies is designed in a V-shape with an aperture 240 at the base of the V through which the spinal therapy device operates. The massage apparatus is arranged with a head support 210 that is integral to the bed (as shown in FIG. 9), although other configurations are also possible. Further support, for example to support the dorsal or lumbar back areas, may also be used (not shown).

[0143] The portion of the bed on which the user lies has an opening or aperture 240 that exposes the back to the manipulating members of the spinal therapy device. This portion of the bed is covered, as shown in FIGS. 8 and 10, such that there is an additional cushioning layer between the manipulating members and the spine, and the user is presented with a flat, bed-like surface on which to lie down. The angle of the aperture 240 with the bed 200, shown in FIG. 10, is designed for comfort and preferably has a value of about 20 degrees to cradle the user's back.

[0144] Dimensions of the bed 200 are preferably 500 mm high x 560 mm wide x 1265 mm long. These dimensions are suitable for transportation and manoeuvrability through doors and passageways. Other dimensions may be used to suit users who are particularly small (children for example) or particularly tall.

[0145] A leg raising portion 220 or cushion is provided, which extends out and may cradle the user's legs using extension mechanism 225. It may be set at a halfway point or fully extended depending on the preference of the user. In some cases, the leg raise may be used to position the user's spine in a particular way, and in this way can form part of the operation of the device in providing a treatment regime.

[0146] The leg raising portion 220 is longer on the top surface of the bed 200 than the bottom surface of the bed 200 to create a comfortable angle for the legs to rest. Preferably wherein parallelogram geometry is created by the top surface being longer than the bottom surface, where the angle at the top corner of the footrest is 8 degrees.

[0147] Reduction of the noise and vibration of the bed during treatment is prevented by spacers 230, which may preferably be made of rubber. The spacers 230 decouple the plate holding the motors from the rest of the chassis 90. In some cases, as shown in FIGS. 1 and 2, the actuators for driving the manipulating members are “floating” relative to the chassis 90. That is to say, they are not mounted on the chassis 90, so do not directly transfer vibrations to the chassis 90. In some cases, mechanical dampers can be included on the transmission path for mechanical vibrations. This can help to further reduce vibration transfer.

[0148] A treatment profile comprises a method to be executed by the spinal therapy bed 200. Individual manipulating members are addressed by the actuator, which causes them to move forwards and backwards in a linear motion. The linear motion produces a force into the user's back in the region of the spine that it engages. This force can be tailored, for example, by the surface area of the manipulating member or the speed with which it is driven. Sufficient force is provided to drive the manipulating member into the user's back with a massaging effect without causing damage.

[0149] Examples of treatment profiles and tailoring may include:

[0150] (1) Area of focus. The amount of time of a treatment profile spent, for example, focussing on a particular area of the spine, e.g. the neck. In some examples, a treatment could spend twice to three times longer on the neck than it does on upper back and/or lower back.

[0151] (2) Speed of the manipulating members: fast movements are felt by a user as being more ‘intense’ compared to slow movements which are perceived to be more ‘gentle’.

[0152] (3) Intensity: ranging from 0 to 10 that represents a fraction of how far each of the manipulating members move toward the corresponding treatment area from a neutral starting position and controls depth of massage into the spine.

[0153] (4) Treatment duration: ranging from around 5 minutes to about 40 minutes, depending on personal needs, time to spare, etc.

[0154] As an example, some specific treatment profiles may focus on the lower back, for example for treating sciatica or vertebral disc problems. Other programmes may be designed for treating headache, arm or shoulder pain caused by trapped nerve(s) around the neck and shoulder, by focussing on the neck and/or upper back region. In this context, “focus” may for example include the manipulating members in the region of focus: spending more time manipulating a particular spinal region; exerting more force while manipulating a particular spinal region; moving a larger distance while manipulating a particular spinal region; and/or moving faster while (relative to the regions which are not regions of focus).

[0155] Two manipulating members may be driven as a pair about a vertebra. The members can be manipulated individually to create a variety of massaging effects. The members can be driven in synchronous motion, for example, by being driven forwards and backwards at the same time to create a squeezing effect. They can also be driven in asynchronous motion such that one moves forwards as the other moves backwards. It will be appreciated that other motions between synchronicity and synchronicity can be performed. Typically, the cycle time of each manipulating member (time taken for a member to move forwards and backwards and return to the starting position) is constant such that the phase of the motion of the members is constant, though it does not have to be. Treatment profiles may comprise a mixture of massaging effects created by the manipulating members in a number of sequences.

[0156] Pairs of members are configured in assemblies such that two pairs address adjacent vertebrae. Members that are adjacent one another along the length of the spine are configured to be driven together. Members in assemblies are therefore typically driven to act as two pairs, so the massaging effect applied to a first pair is simultaneously applied to the second pair.

[0157] Assemblies can be driven with the same massaging effect along the length of the spine or with different effects. A rippling effect along the length of the spine can be developed by driving assemblies with the same cycle length and massaging effect at different starting points in the cycle. Other effects, such as a rocking motion, can be achieved by driving members on each side of the spine asynchronously. Twisting or zig-zag motions can also be created by assemblies in groups at the lumbar and cervical regions being driven in anti-phase with assemblies in the thoracic region. Different regions can experience different massaging effects where it is appropriate for individual users.

[0158] Machine learning techniques analyse data and automate analytical models. It is a branch of artificial intelligence based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention. However, human intervention can be used to overrule machine learning decisions where appropriate. The machine learning can be performed across a large cross-section of users of the spinal therapy bed. Maintaining anonymity can be prioritised, and secure measures implemented to ensure patient confidentiality. However, metadata can be used to very quickly determine effective treatment profiles and learn efficient ways to treat users.

[0159] Parameters of the apparatus and method described above which can be altered in such learning techniques include: spacings and distances between manipulating members in assemblies and in different groupings (for example arranged according to a user's height), force with which to drive the manipulating members, tilt of the manipulating members controlling the angle of manipulation the spinous area of a patient, height of the footrest, treatment profiles including synchronisation of movement of manipulating members in different assemblies and/or on different sides of the spine, total length of treatment or lengths of particular stages or cycles within the treatment, distance of manipulating member to travel in one movement (which may change, or dynamically change, throughout a treatment), operating speed of a manipulating member (which may also change throughout a treatment); motions (e.g. ripple, twist, squeeze, rock) to be performed throughout the treatment and, optionally, instructions on how to achieve these motions using the plurality of manipulating members; duration and sequences of motions; force/pressure to be applied to the spine; and/or other parameters which can be controlled by a central unit or control system. Each of these parameters can be varied by the controller as part of an iterative learning experience. Further input such as measurements of force exerted (via current or voltage draw as set out below); measurement of actual extension as compared with the extension intended to be implemented by the controller; or even user satisfaction (or dissatisfaction) feedback can be used to allow the model to self-identify whether the changes were effective in improving the treatment profile for that particular user. When aggregated, this learning can be extremely useful in guiding the development of highly tailored treatment programs for new or existing users.

[0160] Machine learning and artificial intelligence can be used to improve the user experience and the effectiveness of the treatment. During a treatment, manipulating members move up and down along their axis by an amount determined by a control signal, for example the electrical current imparted by a motor coupled to each manipulating member (or group of members), and a different intensity felt by a user can be achieved based on the range of movement along the axis. This data can be used to model the behaviour of a treatment profile. An example of a model that can be used to determine a treatment profile is detailed below.

[0161] Training can be performed using the following steps. People with pains in different areas of the spine (upper, middle and lower back) are found and participate in a number of sessions until the best settings are found for their specific needs. Whether or not a particular motion has been appropriate can be determined (and fed back into the training model to verify the selection of motions, or suggest that a different motion be tried next time) by any suitable method. For example, users may self report satisfaction with the motion of the manipulating members (either continuously during treatment, or in general at the end of a session, over their whole spine, or in particular areas, etc. —the reporting can be as fine grained as desired). In other cases, the relationship between strength of control signal (e.g. current supplied to a motor) and the force exerted, the distance moved, the speed of motion, etc. can be used to infer whether the treatment provided by the manipulating members was appropriate for the particular tissue being manipulated, on a per-manipulating-member basis.

[0162] Indeed, the machine learning process can even learn from a user over multiple sessions. For example, tissue of a particular stiffness may become gradually relaxed over several sessions. In the training phase, the machine learning algorithm may be informed of the change in tissue stiffness (or other parameter(s)) over time as a user recovers from a particular condition. By inferring the relevant parameter(s) from the dynamics of the manipulating members as discussed above, the machine learning process can infer how far along the treatment is, and cause the manipulating members to make appropriate motions for that stage of the treatment. This process can also be used to feedback to the model, for further refinement, for example where a user is not responding as expected, alternative profiles may be implemented. The success or failure of these for improving the user's condition can be used to guide future treatments for other patients (the data are, of course suitably a.

[0163] Data is collected throughout the sessions concerning the electric currents and individual motor movements. Electrical current for N motors is denoted as X={X1, X2, . . . , XN} and the manipulating member position is denoted as Y={Y1, Y2, . . . , YN}, the AI model is a function Y =f(X). The function, f, is a recurrent neural network model that has the ability to observe the sequence of X over time such that it can make predictions and deliver the most appropriate treatment for a user.

[0164] The model is represented by a set of matrices W and hidden states h. At each time step, the model is given the input electrical current, X, and the output label, Y, then, using a neural network optimisation algorithm (e.g. ADAM, SGD, etc.), the matrices W are adapted to an optimal value.

[0165] Inference can be performed using the following steps. The system takes the electrical currents as input and predicts which part of the body is having a problem and from that, the system can deliver the most suitable treatment program for the user.

[0166] In some embodiments, additional parameters can be monitored, for example to learn when to stop driving the manipulating members into the spine. As the manipulating members push against the tissue, the current reading increases steadily for a short period of time before spiking when they manipulate and deform the tissue and press against the bony mass of the spine. This is illustrated by the sudden change in gradient in the graph of FIG. 11, which plots current against time for one cycle of the movement of the members. The motors driving the manipulating members require more current to push through more dense/higher tension tissue than soft/relaxed tissue. When plotted on an axis of current against time, the point at which the spike appears should decrease with time across the therapy session, indicating that the tissue is loosening. Monitoring the current feedback can be performed and utilised by machine learning techniques. Learning where the threshold or boundary between hard and soft tissue lies can be used in a number of scenarios.

[0167] The point just before a significant increase in gradient, as shown on the graph in FIG. 12 is indicative of a boundary between hard and soft tissue and can be learnt. By learning when in the cycle the change in gradient is going to happen, hitting bony mass can be avoided, which prevents injury or discomfort to the user. These measurements could be used to determine effectiveness of a treatment program and machine learning used to alter the program accordingly.

[0168] Machine learning can be used for personalising individual treatments or for creating more generic treatment profiles. Personalisation can be in one or more of the following areas:

[0169] 1) Treatment area Given electrical current values for each motor, the model can predict the stiffness of the different areas of the back. It is believed that the harder the back, the more electrical current is needed to deliver a constant speed of movement of the manipulating members.

[0170] 2) Treatment intensity and duration

[0171] Given the stiffness predicted, artificial intelligence and machine learning can predict from its training data that a particular stiffness level indicates a certain level of intensity should be used as has been trained from users who participated in training data.

[0172] For example, consider a sample of two users. A first user has a first stiffness of upper back of 8 out of 10 and during a massage, the first user often chooses a level of intensity of 5 out of 7. A second user has a similar stiffness of upper back, for example 7 out of 10, and the second user often chooses the level of intensity of 4 out of 7. This gives the AI training samples: 8:5, 7:4. A generalised model can be built on these results, including a number of additional results, that can be used to determine a treatment given any stiffness level, and any other factors such as users weight for example, and predict an appropriate intensity level for that user.

[0173] Treatment duration follows the same principle as the intensity learning.

[0174] A general library database can be created that stores group data such as applied control settings and measured thresholds, which can be used to inform future treatment profiles. For example, measuring where the hard-soft tissue boundary is likely to occur can be used to pre-set the positioning of physical parameters (e.g. the range of extent of manipulating members, their optimal angle of operation or geometric positioning within an assembly) and operating parameters (e.g. applied force, synchronicity of motion or session length). Group data may be used to, for example, generate informed treatment profiles for new users, which may provide a starting platform that can be personalised as treatment is undergone.

[0175] Treatment settings of individual patients may be stored and saved such that treatment sessions can both be planned according to individual needs and monitored to determine progression made over a series of treatment sessions. User specific details may be stored and accessed from a user library, which may be configured to automatically apply certain features before beginning a new treatment session (such as operating force/power or previously determined control variable thresholds),

[0176] Particular conditions may prove to show particular traits in back stiffness or respond in a predictable way to treatment sessions. Collecting and analysing data from each treatment session may be used to create a series of tailored, selectable treatment profiles, which are continuously modified and improved as more treatments are performed. For example, if it is determined that rippling motions are more successful at treating particularly stiff areas of the back compared to synchronous motions, treatment profiles may be suitably adapted to reflect this.

[0177] Remote diagnosis can also be performed from data collected during treatment, for example, from the current drawn by different components of the system. Such remote diagnosis may be used to select a pre-generated treatment profile from a library of selectable profiles. The library may contain profiles that have been created using machine learning techniques, ready-made profiles, or a mixture of both.

[0178] It will be appreciated from the above description that many features of the different examples are interchangeable with one another. The disclosure extends to further examples comprising features from different examples combined together in ways not specifically mentioned. Indeed, there are many features presented in the above examples and it will be apparent to the skilled person that these may be advantageously combined with one another.