segment kinematics

Rockers or rollers?

Writing about the movement of the hind-foot the a couple of weeks ago and about projection angles last week has led me to reflecting a little on Jacquelin Perry’s rockers. As with many of the concepts that we have in gait analysis, the rockers can give us some really useful insight into how we walk but can also prove misleading if we don’t remain conscious of their limitations.

I don’t recognise the word “rocker” as meaning anything in particular in this context and had assumed it was an American word meaning pivot or fulcrum. I happened to mention this to a couple of American colleagues a couple of years ago, however, and found that they didn’t recognise the word either. It would appear that Perry simply made it up. Not that it matters much, the word seems to get the concepts across readily enough.

The rockers provide mechanisms for the tibia to move forward over the foot and hence for the passenger unit to be carried forward in stance. If we look at the angle the tibia makes to the vertical (above) then we can see that it starts off about 20° behind vertical at foot contact and progresses forwards reasonably steadily (with a bit of a wobble) to reach about 50° in front of vertical at foot off.

tibial progression

Perry explains this in terms of three rockers.  Early on the whole foot rotates about the heel. Later on the tibia rotates over the foot about the ankle and then finally the whole foot rotates about the forefoot (see below). Easy eh!


There is no doubt that all three mechanisms make important contributions to tibial progression. I’m not quite so convinced by Perry’s implication that these occur as a sequence of discrete mechanisms. To investigate this we need to look at the dorsiflexion graph which tells us when ankle rocker occurs and the foot projections graph that tells us when the heel and forefoot rockers are active (see graphs below, note that is impossible to distinguish the timing of the rockers from the ankle angle graph alone ).

Rocker graphs


Heel rocker starts off at foot contact and proceeds until the foot is flat at about 8% of the gait cycle (in red above). It should be noted that this is considerably longer than the period to maximum plantarflexion in early stance that it is sometimes related to. Ankle rocker is the period over which the dorsiflexion angle increases which we can see from the ankle angle graph is from about 5% of the gait cycle to about 45%. There is thus a short period of overlap when both the heel and ankle rockers are active.

Forefoot rocker starts with heel lift which Perry suggests occurs at mid-stance (30% gait cycle). The data depicted above suggests it might commence even earlier (20%?) and it continues until the end of stance. It is thus clear that there is a considerable period from about 20% of the gait cycle until 45% when both ankle and forefoot rockers and simultaneously active.

The conclusion is that whilst the rockers are undoubtedly the mechanisms which allow the tibia to progress they form an overlapping progression rather than a series of discrete events. Indeed for the majority of stance two rockers are active simultaneously.

Since Perry introduced the concepts there has been some slippage in how the terms have been applied which is best avoided. As far as I can see, Perry always talked about heel, ankle and forefoot rockers and never first, second and third rockers. I think this is good practice as quite a lot of our patients don’t have a first rocker (they make contact with the forefoot rather than the heel). It’s always seemed a little illogical to me for someone to have a second rocker if they’ve never had a first rocker!

The other common misconception is that the rockers are alternative labels for phases of the gait cycle. Again Perry never used them in this sense, for her they are mechanisms that allow the tibia to move forward over the foot not phases of the gait cycle. It is particularly erroneous to apply these terms to phases of pathological gait. Many kids with CP never make heel contact and it is thus completely inappropriate to refer to early stance as the phase of heel rocker.

This reinforces the fact that the rockers are mechanisms of normal gait and great care is required in applying the terms to walking with pathology. If a child with CP makes contact with the toe after which the foot comes flat later in stance then they must use a mechanism that might best be described as a reverse forefoot rocker during which the heel is being lowered to the ground rather than being raised. Similarly if they employ a vault to assist clearance of the swing limb then they will often have a reversed ankle rocker during which plantarflexion (rather than dorsiflexion) increases.

Referring back to the work I described in my blog the week before last strongly suggests to me that, in bare feet, the heel rocker is actually a heel roller with the movement being a rolling on the curved surface of the posterior-distal calcaneus rather than a pivot about a particular point on the heel. On the other hand if walking in a shoe with a reasonably stiff heel it is more likely that a rocker like mechanism does occur. The appropriateness of this terminology may thus depend on footwear as well as gait pathology.

PS. In the second edition of Gait AnalysisPerry and Burnfield describe a fourth toe rocker very late in stance.  This can certainly be seen on slow motion videos but I’m not aware of any detailed studies of its biomechanical significance. It looks to occur very late on and I suspect only after most of the load has been taken off the foot but it would be nice to see a more definitive analysis of this.

Projection angles

I’ve had some feedback from Vicon support that people have been asking them how to calculate what I’ve called projection angles on page 138 of my book. These are graphs that look a bit like joint kinematics but represent how each of the segments is aligned with respect to the global axis system rather than to the proximal segment. Two of the femur projection angels thus show how the long axis of the femur is tilted with respect to the vertical in the global sagittal and coronal planes. The third angle shows how the femur is rotated about this axis (projected onto the transverse plane).

projection angles


I first plotted these graphs as a quality assurance tool in that they represent what you should see on  a video recording of the person walking (as long as you take into account parallax effects if the person is not in the centre of the screen or the camera is not directed exactly along one of the principal axes of the global coordinate system). Thus the femur transverse projection tells you whether you should be seeing the femur as internally or externally rotated as viewed by a camera towards which the person is walking. It avoids the need to perform a mental sum of pelvic rotation and hip rotation to assess which is required otherwise. In the example above, at foot contact the left thigh (red) is facing directly ahead and the right thigh is internally rotated by about 5°. You probably won’t be see such a small difference but if the right limb looked to be externally rotated at this instant you might want to question the alignment of thigh markers or knee alignment devices.

Since starting to plot the angles, however, a range of other uses have emerged. The tibia and femur sagittal projections, for example, are essentially what Elaine Owen refers to as segment to vertical angles when tuning ankle foot orthoses.

The foot transverse plane angle is what many of us already plot out routinely and call foot progression. The corresponding angle in the sagittal plane, however, is very rarely plotted but gives a direct appreciation of whether the foot is flat or not. In the example above the foot makes contact at an angle of about 15° to the ground and rotates to become flat on the floor (0°) during about the first 8% of the gait cycle (Perry’s heel rocker). It then remains flat until about 40% of the gait cycle (ankle rocker) after which heel rise causes the foot to start tilting forwards (negative angle, representing toe rocker). If distinguishing between the rockers is important to you then using a graph like this is about the only way to do it. I’ve referred in a previous post to how useful I find this information can be.

I’ve not plotted the pelvic graphs because, if you calculate them using the correct rotation sequence, then they are virtually identical to the pelvic joint angles.

The main reason for this post is thus to make the model that I wrote many years ago to calculate this widely available (click here to go to the download page). Unfortunately it is written in Vicon’s BodyLanguage so will only be directly useful for Vicon users (please note that it requires plugin Gait to have been run first). The accompanying description of exactly what the angles represent should, however, allow any reasonably competent clinical engineer to calculate the equivalents in any other programming/modelling language.